Steen,
A generous offer which I will accept immediately!
This layout should fit on a 150 * 350mm heat-sink easily. I will push it over to the right as far as possible to keep as much of it away from the transformers at the front of the case as possible.
I have used the numbering on the schematic as per my post: http://www.diyaudio.com/forums/showthread.php?postid=1264057#post1264057
Power traces and CCS to Output Mosfet links (I have thinned these on my diagram so you can see the underlying detail) are 5.5mm thick. Signal traces are 1.3mm thick.
1. For my entertainment and education, I have added VR25 and VR33 so I can alter the amount of "Aleph" modulation. (That I why I put the question about Aleph modulation a few posts back.)
2. I have maximized symmetry, using Nelson's suggestion of "folding the right half of the circuit back onto the left half", so that electrical/magnetic fields will be as equal as possible on the right and left halves.
3. I have added a jumper to join Signal Negative to Ground for single ended input.
4. I have added an array for input capacitor options.
5. I have added zener protection to a 3 ohm resistor (R0) to earth, as per the Aleph-X thread.
6. The "box" resistors are 0.4" spacing for Vishay bulk foil resistors, if I ever want to put them in.
7. I will be "Toner Transferring" with my wife's iron - Choky always likes that, so there are a few odd vias which would not be necessary with a professional board.
8. There will be a ground plane over the low voltage end of the board, or a thick "star" ground arrangement at that end as a top layer.
Unfortunately, I cannot encourage Circad 98 to show the names of the signal nets or pads on the printout ... sorry.
If I need to parallel the 2SJ109 or cascode it (to get more oomph to the mosfets ... they are IRFP150N, which are approximately one-and-a-half 140N's each, I will put a daughter board above this one. (But I will need lots of advice on how to implement a cascode that. There was lots of talk about it in the Aleph-X thread, but then the thinking wandered off to XA topologies.)
(Unfortunately as a zip file, because it is unreadable if the picture is less than 1000 pixels wide.)
Extreme thanks for your input,
Regards,
George.
GeorgeBoles said:
it's nice to see that pcb ;
but- if I'm lousy in anything (
) , I'm lousy in pcb checking ....... so - I'll let Steen and others to check and comment .......on the other side - if you need few rusty ideas how to cascode tiny critters , mebbe I can help with that
even if I can't see any necessity of cascodes if PS is bellow say - 2 x 25V
GeorgeBoles said:
George, that all sounds like good reasoning.
With the 15V rails in the stock Babbel-X, it is not needed to cascode the input pair, so I suggest that you use the schematic as it is, for now. I will make a print of both the schematic and your boardlayout, and look it over this weekend.
Steen
Rails ...
Thanks Steen,
I have 18V secondary transformers ... and I am going to build with these.
This will produce rails of 24 volts or a little less if I add some resistance to the CRC or put in a slightly high resistance L in a CLC.
Regarding cascoding:
Choky-Zenmod had warned us that the 2SJ109 will not drive a large number of MOSFETS, and he made that statement about the Babelfish-J, which only has TWO mosfets to drive. I thought that driving FOUR in the Babelfish-J-X might be overdoing it a bit. If the IRFP150N was like a 50% bigger IRFP140/240N, then it really might not work that well. That WAS what I was worried about mainly.
The next problem is that to get a generous bias current (to get "good" sound), with the higher voltage rails, I might cook the IRFP150N. My calculations indicate I will be dissipating about 31W per device with my current plan. If that is too much, I will need to parallel my MOSFETs ... and then there will be 8 mosfets to drive, which will be FAR too many! And therefore my worry about the ability of the 2SJ109BL to drive the output devices.
I saw Nelson suggest "just parallel the input device" to solve this problem, which is something I can understand. Other contributors have talked about cascoding as another way of dealing with the 2SJ109. Will cascoding also provide more "driving power"?
That is the reasoning in my letter. I hope that makes sense.
I am itching to go and burn some circuit boards ... itch, itch, itch.
Best wishes to all,
George
P.S. The gate capacitances are not that different:
1400pF 140n
1900pF 150n
1300pF 240N
which is reassuring.
Thanks Steen,
I have 18V secondary transformers ... and I am going to build with these.
This will produce rails of 24 volts or a little less if I add some resistance to the CRC or put in a slightly high resistance L in a CLC.
Regarding cascoding:
Choky-Zenmod had warned us that the 2SJ109 will not drive a large number of MOSFETS, and he made that statement about the Babelfish-J, which only has TWO mosfets to drive. I thought that driving FOUR in the Babelfish-J-X might be overdoing it a bit. If the IRFP150N was like a 50% bigger IRFP140/240N, then it really might not work that well. That WAS what I was worried about mainly.
The next problem is that to get a generous bias current (to get "good" sound), with the higher voltage rails, I might cook the IRFP150N. My calculations indicate I will be dissipating about 31W per device with my current plan. If that is too much, I will need to parallel my MOSFETs ... and then there will be 8 mosfets to drive, which will be FAR too many! And therefore my worry about the ability of the 2SJ109BL to drive the output devices.
I saw Nelson suggest "just parallel the input device" to solve this problem, which is something I can understand. Other contributors have talked about cascoding as another way of dealing with the 2SJ109. Will cascoding also provide more "driving power"?
That is the reasoning in my letter. I hope that makes sense.
I am itching to go and burn some circuit boards ... itch, itch, itch.
Best wishes to all,
George
P.S. The gate capacitances are not that different:
1400pF 140n
1900pF 150n
1300pF 240N
which is reassuring.
Circad 98
By the way, is it just Grey Rollins and myself who are the only people on the board who use Circad98, demo version?
It is unrestricted for private use, the only thing being that it doesn't produce Gerber files unless you pay for a licence. Occasionally it does a strange thing changing the label of Top Copper to Top Silk, and occasionally it changes the grid snap to a coarser setting, but really it has been very good for me! In fact, Circad and my wife's iron got me into DIY electronics!
Oh, well - that is just, as I said, by the way ...
Regards,
George
By the way, is it just Grey Rollins and myself who are the only people on the board who use Circad98, demo version?
It is unrestricted for private use, the only thing being that it doesn't produce Gerber files unless you pay for a licence. Occasionally it does a strange thing changing the label of Top Copper to Top Silk, and occasionally it changes the grid snap to a coarser setting, but really it has been very good for me! In fact, Circad and my wife's iron got me into DIY electronics!
Oh, well - that is just, as I said, by the way ...
Regards,
George
Re: Rails ...
Babelfish J - 2SJ109 is driving one gate ; you can safely drive two
each Mosfet is burnin' at 50W without problems
cascoding can save your eeny weeny 7 legged critter from excessive vvoltage and dissipation,but you can't increase current through them without sense and without taking in account other param's as Ugs ,etc.
so- leave it at ~ 10mA through both halves (5mA each half ) and enjoy
you can also unwound little those donuts ; it's easier than you think and more elegant path- stay at DIY and make just most effective compromises.....
even if you must choose and buy another set of donuts- you really don't build one of best world's amps every day?
"watts ?
who need watts?!! "
GeorgeBoles said:
Babelfish J - 2SJ109 is driving one gate ; you can safely drive two
each Mosfet is burnin' at 50W without problems
cascoding can save your eeny weeny 7 legged critter from excessive vvoltage and dissipation,but you can't increase current through them without sense and without taking in account other param's as Ugs ,etc.
so- leave it at ~ 10mA through both halves (5mA each half ) and enjoy
you can also unwound little those donuts ; it's easier than you think and more elegant path- stay at DIY and make just most effective compromises.....
even if you must choose and buy another set of donuts- you really don't build one of best world's amps every day?
"watts ?
who need watts?!! "
I am not all that fond of rewinding trafo's. As appealing as it sounds, I find it more troublesome than good is! It takes a lot of work to unwind the kapton tape, and unwinding the copper is always a bit of guesswork and then you have to wind the kapton back on!! I suggest that you implement a regulated supply instead It burns off the excess voltage, and disposses of any hum from the supply in the process. Take a look at the Zen article #3 (I think) regarding active supply regulation! Do pay attention though regarding the dualrail supply!! The negative mosfet is wrong!!
Steen
It burns off the excess voltage, and disposses of any hum from the supply in the process. Take a look at the Zen article #3 (I think) regarding active supply regulation! Do pay attention though regarding the dualrail supply!! The negative mosfet is wrong!! Steen
Thanks for useful thoughts, everyone.
Zenmod -
Of course! I have counted the number of MOSFETs being driven incorrectly. What a dummy! I can relax a bit now.
I am not sure if I understand this fully: If I needed to, would cascoding, say with a bigger BJT, increase the ability to drive more MOSFETs - for a much larger amplifier, say for example 60 or 100W??
MPmarino -
I really like your idea of putting some "wrong direction" turns on the transformer! I had already thought about some extra windings for LED's or other low voltage tasks - 5V control circuits or similar. I suppose that a regulator might be easier, though.
Steen-
Everyone-
Regarding getting more drive current through the 2SJ109, why has Nelson's pearl about "simply paralleling" them vanished into the ether (i.e. appear to have been forgotten)?
Thanks for your help,
Regards,
George.
Zenmod -
Of course! I have counted the number of MOSFETs being driven incorrectly. What a dummy! I can relax a bit now.
I am not sure if I understand this fully: If I needed to, would cascoding, say with a bigger BJT, increase the ability to drive more MOSFETs - for a much larger amplifier, say for example 60 or 100W??
MPmarino -
I really like your idea of putting some "wrong direction" turns on the transformer! I had already thought about some extra windings for LED's or other low voltage tasks - 5V control circuits or similar. I suppose that a regulator might be easier, though.
Steen-
I presume here you are talking about the Zen 3 article, not my circuit board - have you had time to look at that yet?
Everyone-
Regarding getting more drive current through the 2SJ109, why has Nelson's pearl about "simply paralleling" them vanished into the ether (i.e. appear to have been forgotten)?
Thanks for your help,
Regards,
George.
steenoe said:I am not all that fond of rewinding trafo's. As appealing as it sounds, I find it more troublesome than good is! It takes a lot of work to unwind the kapton tape, and unwinding the copper is always a bit of guesswork and then you have to wind the kapton back on!! I suggest that you implement a regulated supply instead
Steen
Lazy Luke .
GeorgeBoles said:
I wrote some numbers in quote of yours ;
so:
1. with cascoding- main benefit is that shield (preserve) jfets (in that case; in other case some other cascoded part) of excessive PS voltage ..... so you can use much higher PS voltage than without cascoding ; you still can't use more mA through cascode ,because entire current is still passing through tiny critter.....there is another trick Papa use for augmenting stage current ability ,different than paralleling devices ,but- according to him - that trick is much more prominent in output stages...... so- look at No. 4 part of this Re
2. just be sure that you use same wire thickness
for that bucking windings3. certainly Luke is talking about Zen related schm mistake ;
4 . it's not forgotten .
what's No. of output devices you need ,anyway ? greedy for watts ?
you can easily make Bab X with two pairs of IRFP150 on each side (ie. each half of 2SJ109BL is driving only two gates) and have entire dissipation in range of 200W/ch. without much hassle about their (IRFP's) longevity ...... but- is it sane ? or- is it not enough ?
if not- build bigger amp , not Bab X .
Thanks, Zenmoddy,
Exactly what I needed to know, especially about the cascoding. I am a happy camper now.
No, I am not greedy for Watts at present. I have a "lifetime's supply" (well a few anyway) of 2SJ109BL and I am continuously checking where I can use them in another project! And, if I ever come across a set of B&W 802 Diamonds, then I am going to need a BIG amp to push them along ... you never can tell what my "dearest" will allow me to do.
I am trying to understand the information on X and XA in the Aleph-X thread, which may be more useful for "big" amps for "big" speakers ... but I have to get the Babelfish-JX going first. I am just waiting for Steenoe's expert opinion on my second-ever PCB effort, and collecting switches, screws, bolts and bits of aluminium for the construction.
Regards,
George.
Exactly what I needed to know, especially about the cascoding. I am a happy camper now.
No, I am not greedy for Watts at present. I have a "lifetime's supply" (well a few anyway) of 2SJ109BL and I am continuously checking where I can use them in another project! And, if I ever come across a set of B&W 802 Diamonds, then I am going to need a BIG amp to push them along ... you never can tell what my "dearest" will allow me to do.
I am trying to understand the information on X and XA in the Aleph-X thread, which may be more useful for "big" amps for "big" speakers ... but I have to get the Babelfish-JX going first.
I am just waiting for Steenoe's expert opinion on my second-ever PCB effort, and collecting switches, screws, bolts and bits of aluminium for the construction.Regards,
George.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.