Mr Børresen,
it was just a response to the claim that a BD140 is a not well regarded transistor.
A Phily-16 is a Philips BD140-16, but a totally different animal than say a Fairchild BD140.
Old news, told many times before at this forum, by experienced folks as Mr Alex Nikitin, by me.
it was just a response to the claim that a BD140 is a not well regarded transistor.
A Phily-16 is a Philips BD140-16, but a totally different animal than say a Fairchild BD140.
Old news, told many times before at this forum, by experienced folks as Mr Alex Nikitin, by me.
Attachments
Last edited:
Agree that the Phillips have seemingly way higher Ft than so called
equivalents from other manufacturers.
Though, the best results in this schematic are obtained by more
or less recent BJTs like the 2SA1360..
Will do a sim with Phillips models of BD140...
In the waiting, a comparison of mosfet and Bjt VAS versions,
using Lineup s schematic.
Attachments
BJT is exponential transfer; mosfet is square law.
Distortion profiles will be different, with the mosfet having less higher order artefacts.
THD will indeed be lower for the bjt, but what there is will be more sonically objectionable.
However, if THD is your goal, Wahab is absolutely right.
Can you hear the difference, that is the question?
Yes, but only if you build and listen, and even then you need to listen in controlled circumstances with a large sample of both genres and ears.
BTW Stanley, substituting a 1.7A mosfet for a 74A device in the VAS position is fraught with issues, mostly to do with the huge gate capacitance which will slow the amplifier hugely, giving you appalling square waves!! Best to use the IRF9610, which has a small gate of just 170pF Ciss, easily driven by a jfet at 4mA.
Hugh
Distortion profiles will be different, with the mosfet having less higher order artefacts.
THD will indeed be lower for the bjt, but what there is will be more sonically objectionable.
However, if THD is your goal, Wahab is absolutely right.
Can you hear the difference, that is the question?
Yes, but only if you build and listen, and even then you need to listen in controlled circumstances with a large sample of both genres and ears.
BTW Stanley, substituting a 1.7A mosfet for a 74A device in the VAS position is fraught with issues, mostly to do with the huge gate capacitance which will slow the amplifier hugely, giving you appalling square waves!! Best to use the IRF9610, which has a small gate of just 170pF Ciss, easily driven by a jfet at 4mA.
Hugh
Last edited:
Jacco,
I built the amp with a jfet input and BD140-16 VAS. Thought it sounded terrible. I presented the possibility that a different VAS transistor might sound better because I know of some people, who's opinions I respect, who think it is an inferior transistor. It is not my personal opinion and hence why I have used BC550C/560Cs and BD139-16/140-16s for every amplifier I have built.
Personally I think the BD140 VAS sounded terrific with the mosfet input stage, and I made this plainly clear many many posts ago. But it sounded terrible with the jfet input stage. I don't exactly know why, maybe something to do with the load the VAS presents, maybe something to do with the low gain of the jfet.
If you want to talk simulated figures, as some do, the circuit actually simulates better with a mosfet VAS. Not that that means anything, right?
Now, as the only person I am aware of who has built all these circuits, my sonic tastes might be questionable and not in agreement with the majority, or even a single other person. I'm perfectly happy for someone to physically build another version of the circuit with BJTs and tell me this is so.
I built the amp with a jfet input and BD140-16 VAS. Thought it sounded terrible. I presented the possibility that a different VAS transistor might sound better because I know of some people, who's opinions I respect, who think it is an inferior transistor. It is not my personal opinion and hence why I have used BC550C/560Cs and BD139-16/140-16s for every amplifier I have built.
Personally I think the BD140 VAS sounded terrific with the mosfet input stage, and I made this plainly clear many many posts ago. But it sounded terrible with the jfet input stage. I don't exactly know why, maybe something to do with the load the VAS presents, maybe something to do with the low gain of the jfet.
If you want to talk simulated figures, as some do, the circuit actually simulates better with a mosfet VAS. Not that that means anything, right?
Now, as the only person I am aware of who has built all these circuits, my sonic tastes might be questionable and not in agreement with the majority, or even a single other person. I'm perfectly happy for someone to physically build another version of the circuit with BJTs and tell me this is so.
This is something I'll have to try when I can. Unfortunately I have no Jfets and the only MOSFETs I can use are IRF250's. 2nF!!!
Both of those are running 800mA right now, for lack of a suitable resistor larger than 10R, 10W. The image "pops out", but the treble is harsh and fatiguing. I may try lowering the 1kR gate stoppers!!!
- keantoken
Both of those are running 800mA right now, for lack of a suitable resistor larger than 10R, 10W. The image "pops out", but the treble is harsh and fatiguing. I may try lowering the 1kR gate stoppers!!!
- keantoken
MiiB,
from my angle, it is similarly disturbing to watch people persistently advocating FETs for voltage amplification. Some reading would be useful on current and voltage sources in general in conjunction with amplification device properties and function in different topologies. The resistive nature (and other characteristics) of FETs makes them superior transconductors, but that´s a disadvantage for a voltage source. Large fast moving voltage swings don´t get on well in such hindering and lossy environment so the voltage gain and the operating area of FETs should be kept small. Bipolars have many annoying features, but FETs can impossibly be equally good at carrying out two opposite tasks.
Simulations won`t tell you anything of value in this regard either.
from my angle, it is similarly disturbing to watch people persistently advocating FETs for voltage amplification. Some reading would be useful on current and voltage sources in general in conjunction with amplification device properties and function in different topologies. The resistive nature (and other characteristics) of FETs makes them superior transconductors, but that´s a disadvantage for a voltage source. Large fast moving voltage swings don´t get on well in such hindering and lossy environment so the voltage gain and the operating area of FETs should be kept small. Bipolars have many annoying features, but FETs can impossibly be equally good at carrying out two opposite tasks.
Simulations won`t tell you anything of value in this regard either.
As i see it ...Hugh is of one opinion weighting the distortion profile and you is of another....The BJT may be the most suitable VAS device, but if the the combination with the Jfet does not work well....then all is in Vain....!!
Any way.. there seem to be no correlation between what we musically like and to the level of distortion.. Hugh may be closer to the truth.. It's not so much the level as it is the distribution that matters...but to go to the extent of injecting H2-H4 buy using a bootstrap is not so much my cup of tea..
This is one reason why I belive my version with the BJT VAS and Hawksford cascode will perform quite well...as it seems to keep the distribution regardless of signal levels..
I'am by no means an EE. my science background is materials and vibration wich works quite well with loudspeakers and drive units, but leaves me quite helpless with the theoretical background in this game....And I do really appreciate when people here takes the time to give sound explanations. I believe it gives value and credit to the forum.. Think there a big difference in understanding device basics and understanding circuit specific applications, selections and working conditions...hope this deeper circuit understanding will come with time...
Any way.. there seem to be no correlation between what we musically like and to the level of distortion.. Hugh may be closer to the truth.. It's not so much the level as it is the distribution that matters...but to go to the extent of injecting H2-H4 buy using a bootstrap is not so much my cup of tea..
This is one reason why I belive my version with the BJT VAS and Hawksford cascode will perform quite well...as it seems to keep the distribution regardless of signal levels..
I'am by no means an EE. my science background is materials and vibration wich works quite well with loudspeakers and drive units, but leaves me quite helpless with the theoretical background in this game....And I do really appreciate when people here takes the time to give sound explanations. I believe it gives value and credit to the forum.. Think there a big difference in understanding device basics and understanding circuit specific applications, selections and working conditions...hope this deeper circuit understanding will come with time...
Add the "magic resistor" between VAS active devices and cascode you can alter early effect and change hugh's "balance". Miib is correct , the hawksford keeps it's distortion profile from subsonics to ultrasonics. I'm trying mosfet's for the cascode and the results are promising (no offset issues). You do lose a lot of voltage swing with Vgs, but I use boosted rails anyway.All the OEM's use the cascoded fet pair for LTP. Possibly , being voltage controlled and square law based, it has a different effect on the global NFB loop. The output stage and reactive speaker load have less effect on either a CFP BJT or fet LTP. I think this is the "magic" (soundwise) as far as the IPS is concerned.
"FET's for voltage amplification"... better as cascodes in the VAS . Also better than a plain bjt for the IPS.
I'm trying everything now - nothing is "too crazy" anymore.
OS
I take something who has actually built and listened to the amplifier more seriously than any other opinion.
This is why Fetzilla I recommend has now 2SK170 JFET input and IRF9610 VAS.
Because this was my original idea and what sounded good to swordfishy.
In spice, with the combination of Jfet i/p and BJT VAS - bd140, this amp only had OLG of 70dB - 26dB CLG = 45dB FB - not very much
An identical circuit the irf9610 as VAS gave an OLG of well over 100dB and it rolls off nicely so this amp a can operate without any compensation with a resistive load and produce a very creditable 100KHz square wave, has very low THD with a nice signature, and swordfishy says it sounds great. Good grief, what more do you want !
So for me any theoretical reasons why FET VAS don't work well are really swept away into total insignificance.
If you are desparate to use a BJT VAS a cascode with 2 BD140 ( or similar ) does work quite well and gives similar THD to the irf9610 with a nice signature with less OLG ( just under 100dB ) but, it is more complicated and needs more attention to stabilisation.
mike
An identical circuit the irf9610 as VAS gave an OLG of well over 100dB and it rolls off nicely so this amp a can operate without any compensation with a resistive load and produce a very creditable 100KHz square wave, has very low THD with a nice signature, and swordfishy says it sounds great. Good grief, what more do you want !
So for me any theoretical reasons why FET VAS don't work well are really swept away into total insignificance.
If you are desparate to use a BJT VAS a cascode with 2 BD140 ( or similar ) does work quite well and gives similar THD to the irf9610 with a nice signature with less OLG ( just under 100dB ) but, it is more complicated and needs more attention to stabilisation.
mike
Last edited:
I'm going to have to make some simple headphone circuits to test out different transistors in various positions, and see what differences I can hear. I wonder what's the difference between Phillips and Fairchild BC550/560? Are the Fairchild imitations really that sloppy? I know that the Phillips datasheets show clearly different specs. However I prefer the robust Hfe of the Fairchild datasheets.
Well, I remember that I bought some Fairchild MPSA18 a while back. The TO92 cases were put together with both halves askew. That was a big put-off to me.
I suppose what I really SHOULD do is construct a simple headphone amp with the three monumental positions - input, VAS, output transistors, and then servo it so I don't have to worry about offset. Then I can make every resistor a trimmer. That would allow me to plug & dial. Where do I get transistor sockets?
- keantoken
Well, I remember that I bought some Fairchild MPSA18 a while back. The TO92 cases were put together with both halves askew. That was a big put-off to me.
I suppose what I really SHOULD do is construct a simple headphone amp with the three monumental positions - input, VAS, output transistors, and then servo it so I don't have to worry about offset. Then I can make every resistor a trimmer. That would allow me to plug & dial. Where do I get transistor sockets?
- keantoken
I would tag along with LTSpice, but I can't tell what we're simulating to determine. There's nothing left but to audition as far as I can tell. If we want to reach a consensus the best thing to do is to pass along our prototypes in the mail for auditioning, and let the diverging philosophies go their separate ways, rather than busting square pegs into round holes.
BTW, I really like the musical prototypes through the mail idea, I think it would make some decisions much clearer. But I am not speaking from experience...
- keantoken
BTW, I really like the musical prototypes through the mail idea, I think it would make some decisions much clearer. But I am not speaking from experience...
- keantoken
why?
if you have no information on amp build quality, listening environment, source, speaker/room quality?
wouldn't you at least like to know how they score on Audio Musings by Sean Olive: Harman's "How to Listen" Listener Training Software Now Available as Beta
I wouldn't give any weight to "naive subjectivist" listening impressions if they don't have 2 complete circuits to switch rapidly, level match, and a blinding protocol
audio memory is poor, expectation effect large and too much promotion has gone into audio "connoisseurism" for you to not expect random commentary over the web is someone trying to ape the industry rags like Stereophile reviewer's commentary
Mike,
Opinions are like any other part of the body. Everyone has them. Some are a breath of fresh air, some are malodorous.
The forum offers us an opportunity to vent our thoughts, express ourselves, assert our individuality. Insofar as most people carry some mental illness, the results are variable. Forums are also entertainment, though timing and nuance is poor, and often people misread the motives of others. My opinions are nothing special, I offer what I have found works for me. When you add to the mix the fact that people's taste in audio is diverse, what designs are good and what are bad is often a simple matter of choice.
I am reluctant to make categorical statements, like 'BJTs are current driven devices'. In fact, most engineering minds are reductionist and try to distill phenomenon to a simplified essence, a necessary part of most engineering pursuits, like building a bridge, or designing an engine. But audio is more like cooking, unfortunately, not so amendable.
I don't want to mislead the thread. Some of my ideas are dead wrong, but in truth I'm my own worst critic.
You picks your poison and runs with it......
Hugh
Opinions are like any other part of the body. Everyone has them. Some are a breath of fresh air, some are malodorous.
The forum offers us an opportunity to vent our thoughts, express ourselves, assert our individuality. Insofar as most people carry some mental illness, the results are variable. Forums are also entertainment, though timing and nuance is poor, and often people misread the motives of others. My opinions are nothing special, I offer what I have found works for me. When you add to the mix the fact that people's taste in audio is diverse, what designs are good and what are bad is often a simple matter of choice.
I am reluctant to make categorical statements, like 'BJTs are current driven devices'. In fact, most engineering minds are reductionist and try to distill phenomenon to a simplified essence, a necessary part of most engineering pursuits, like building a bridge, or designing an engine. But audio is more like cooking, unfortunately, not so amendable.
I don't want to mislead the thread. Some of my ideas are dead wrong, but in truth I'm my own worst critic.
You picks your poison and runs with it......
Hugh
Hi Hugh,
not sure what you are trying to say here !
are you hinting that you are coming up with a stonkingly good version of this amp with a BJT VAS ? if so - great !
knock me down with a feather
I'm just cautious about believing "theories" of why an amp that measures and sounds great is somehow wrong.
not sure what you are trying to say here !
are you hinting that you are coming up with a stonkingly good version of this amp with a BJT VAS ? if so - great !
knock me down with a feather
I'm just cautious about believing "theories" of why an amp that measures and sounds great is somehow wrong.
My simulations using IRF9610 are also good results.
So, I do not only recommend IRF9610 by listening impression.
I am glad you, mikelm, have found same as me: IRF9610 works very well.
If every amplifier would use only THD and technically 'best' devices
in every place, like in VAS,
then we would have only one way to make THE AMPLIFIER.
Fetzilla is different.
This is the good thing.
Otherwise it would be just another blameless.
And Fetzilla has proven to work and also sound good.
No, Mike, I have nothing better than SWF has produced.
But I am a bit concerned about the 7V spike at switch on. I know it could be worse, but it could be better, too, and just wondering if DC coupling with a servo might actually improve things.....
My only concern with distortion is the higher orders, from H5 on. They MUST be minimised, and if H2/H3/H4 is still high, that is generally no problem.
Cheers,
Hugh
But I am a bit concerned about the 7V spike at switch on. I know it could be worse, but it could be better, too, and just wondering if DC coupling with a servo might actually improve things.....
My only concern with distortion is the higher orders, from H5 on. They MUST be minimised, and if H2/H3/H4 is still high, that is generally no problem.
Cheers,
Hugh
not because sims are "the last word" but because its easy
it is entirely possible for some propositions about circuit's audibility made in this thread to be tested by running music .wav thru the simmed topologies and using Foobar or similar ABX plugin
with high rez samples, good DAC and decent headphone's the "resolution" should be quite high - higher than most loudspeaker/room setups for some "listening dimensions", particularly distortion
can anyone really tell what (simmed) VAS is being used?
it is entirely possible for some propositions about circuit's audibility made in this thread to be tested by running music .wav thru the simmed topologies and using Foobar or similar ABX plugin
with high rez samples, good DAC and decent headphone's the "resolution" should be quite high - higher than most loudspeaker/room setups for some "listening dimensions", particularly distortion
can anyone really tell what (simmed) VAS is being used?
Last edited:
Yes, I must apologise for having to live with my audio setup in an ordinary house rather than an anechoic chamber, for not having 5 sets of $100k speakers and amplifiers in a setup which allows easy switching, and for not having a team of ready participants for a blind A/B test. I must also apologise for being tired on some days, mad on others and happy on even more, and for letting that influence my impressions.
I must apologise for building perhaps 10 incarnations of this circuit with an open mind, for trying mosfets, jfets and BJTs in most of the critical parts of the circuit, for listening for hours and finally deciding on the circuit which I had been most against at the outset.
And finally, I must apologise that the circuit I decided on, unbeknown to me, also happened to have the most favourable simulation character.
Yes, my opinion means nothing. Maybe someone else will build the amplifier too and prove me wrong, and maybe they will have the anechoic chamber and ideal listening conditions too. Just like all of us hobbyists who are just doing this for the fun and passion of it have.
Seriously though, someone else please build it and let us know what you think.
In the meantime I'm going to enjoy building simple circuits that sound nice to my ears and conversing about it with nice open people on the forum. I also hope the guy who bought my Doug Self blameless amplifier is enjoying its miniscule distortion, high level of detail and lack of soul.
I must apologise for building perhaps 10 incarnations of this circuit with an open mind, for trying mosfets, jfets and BJTs in most of the critical parts of the circuit, for listening for hours and finally deciding on the circuit which I had been most against at the outset.
And finally, I must apologise that the circuit I decided on, unbeknown to me, also happened to have the most favourable simulation character.
Yes, my opinion means nothing. Maybe someone else will build the amplifier too and prove me wrong, and maybe they will have the anechoic chamber and ideal listening conditions too. Just like all of us hobbyists who are just doing this for the fun and passion of it have.
Seriously though, someone else please build it and let us know what you think.
In the meantime I'm going to enjoy building simple circuits that sound nice to my ears and conversing about it with nice open people on the forum. I also hope the guy who bought my Doug Self blameless amplifier is enjoying its miniscule distortion, high level of detail and lack of soul.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.