bjt version has very good thermal stability. While the mosfet drives at about 10-20mv (on the 0.20ohm source resistors), the bjt's voltage drop off the 0.47ohm emitter resistors is about 1-2mv.
I think the current sharing between the upper and lower output devices, while not exactly equal, helps stabilit Iq. the MOSFet version, being voltage driven, doesn't have that benefit.
I think the current sharing between the upper and lower output devices, while not exactly equal, helps stabilit Iq. the MOSFet version, being voltage driven, doesn't have that benefit.
Tschrama, I sent you a PM concerning your new heatsink. I see you're away a fortnight. When you react to my mail with your adress I can send the thing....
Happy 2004 everybody !!
Happy 2004 everybody !!
😀 😀 YES, BY ALL MEANS A HAPPY 2004 TO EVERYBODY, HICC!
AS YOU CAN SEE Í've already had my celebration as I am 7.5 hrs ahead of most of you where I am,,,, so 1/2 a bottle of vodka down another 1/2 to go, pity it isnt stolichnaya just a cheap NZ made volska ohhhh welll can't have everything now can I?
ohhh, my poor liver!, not to mention my pancreas...
cheers to all an shundry
Tomcat:at least that's who I think I am at this moment xeye: perhaps by the end of the day I Wished I was somebody else yeay for good ol'' john linsley hood...
AS YOU CAN SEE Í've already had my celebration as I am 7.5 hrs ahead of most of you where I am,,,, so 1/2 a bottle of vodka down another 1/2 to go, pity it isnt stolichnaya just a cheap NZ made volska ohhhh welll can't have everything now can I?
ohhh, my poor liver!, not to mention my pancreas...cheers to all an shundry
Tomcat:at least that's who I think I am at this moment xeye: perhaps by the end of the day I Wished I was somebody else yeay for good ol'' john linsley hood...I did a listening test between:
a) the JLh design vs. the Graham modification where the 8.2k resistor on the collector goes to the emitter resistor on the driver transistor; and
b) MOSFET driver (a fairchild medium power mosfet) vs. BJT driver (Toshiba 2sc4793); and
c) MOSFET output devices (IRF540x2) vs. BJT output devices (TIP41Cx2). Iq set at 750ma as that's the Iq for the BJT version where I cannot adjust Iq easily.
The switch is a Parasound preamp (p/hp-850). I ran the highly unscientific tests about 20 runs on a) and b).
c), being the easiest one, was decided after a few songs: I thought that the mosfet was "cleaner", especially in the low-end. Granted, this isn't as realistic a test as no one would build the BJT version with the TIPs. so I will do a test pitting the IRFs against MJE15030s later.
on a) and b), I simply couldn't tell the two apart. I would give a marginal, probably imaginery, advantage to the 2SC4793, as it seemed a little bit finer on the high-end of the spectrum.
the Graham modification worked very well in simulations for closed loop performance but seems to have narrower open loop frequency response, which is why I decided against it.
Well, happy new year to everyone.
a) the JLh design vs. the Graham modification where the 8.2k resistor on the collector goes to the emitter resistor on the driver transistor; and
b) MOSFET driver (a fairchild medium power mosfet) vs. BJT driver (Toshiba 2sc4793); and
c) MOSFET output devices (IRF540x2) vs. BJT output devices (TIP41Cx2). Iq set at 750ma as that's the Iq for the BJT version where I cannot adjust Iq easily.
The switch is a Parasound preamp (p/hp-850). I ran the highly unscientific tests about 20 runs on a) and b).
c), being the easiest one, was decided after a few songs: I thought that the mosfet was "cleaner", especially in the low-end. Granted, this isn't as realistic a test as no one would build the BJT version with the TIPs. so I will do a test pitting the IRFs against MJE15030s later.
on a) and b), I simply couldn't tell the two apart. I would give a marginal, probably imaginery, advantage to the 2SC4793, as it seemed a little bit finer on the high-end of the spectrum.
the Graham modification worked very well in simulations for closed loop performance but seems to have narrower open loop frequency response, which is why I decided against it.
Well, happy new year to everyone.
Talkin' to yourself again?
Millwood have you noticed you are talking to yourself? Being chatty and wishing everyone a Happy New Year does not make up for your rudeness and bad behaviour of earlier posts. Apologize to Graham and perhaps Geoff and we can get them back online with their learned input and not just your over-sensitive-to-criticism reactionary posts and masturbatory simulations.
Millwood have you noticed you are talking to yourself? Being chatty and wishing everyone a Happy New Year does not make up for your rudeness and bad behaviour of earlier posts. Apologize to Graham and perhaps Geoff and we can get them back online with their learned input and not just your over-sensitive-to-criticism reactionary posts and masturbatory simulations.
thank you Greg for your kind suggestions but rest assured none will be taken.
Happy new year to you, bad boy, 🙂
Happy new year to you, bad boy, 🙂
😀 AHHEM! I THINK GRINGO AUDIO IS WRONG, IF ANYONE IS RUDE AND SELF MASTURBATORY...IT IS HIMSELF I HAVE BEEN FOLLOWING THIS THREAD FOR SOME TIME AND THERE NEVER WAS ANY HINT OF MALICE BY MILLWOOD TO ANY OF THE OTHER CORRESPONDENTS JUST A EUPHORIC SHARING OF DISCOVERIES, WHICH IS WHAT THIS THREADIS ALL ABOUT...... 
TC
TC
Hi Magyarian (qwad),
I second your preference for Stolichnaya. The best Vodka on Mother Earth. Also the cheapest. Thank God for the Ruskies !
The only great Vodka you can drink neat and not burn your throat ( and other vital parts ). The Ruskies taught me this and now I believe it- after considerable practice and experimentation with other brands.
My wish for your New Year-- may you get several bottles of Sotlichnaya to see you through 2004 ! Me too ! Me get Smirnoff - just a pale duplicate!
By the way are you an Alien ? - you know UFO etc .....
Magyar sure sounds like it is from Mars or similar regions. Anyway Happy New Year to all Earthiens and Aliens ! Do Aliens have ears ? Might be broadband .
yeah, I am still guzzling the bottle and having a great time.
Happy New Year.
I second your preference for Stolichnaya. The best Vodka on Mother Earth. Also the cheapest. Thank God for the Ruskies !
The only great Vodka you can drink neat and not burn your throat ( and other vital parts ). The Ruskies taught me this and now I believe it- after considerable practice and experimentation with other brands.
My wish for your New Year-- may you get several bottles of Sotlichnaya to see you through 2004 ! Me too ! Me get Smirnoff - just a pale duplicate!
By the way are you an Alien ? - you know UFO etc .....
Magyar sure sounds like it is from Mars or similar regions. Anyway Happy New Year to all Earthiens and Aliens ! Do Aliens have ears ? Might be broadband .
yeah, I am still guzzling the bottle and having a great time.
Happy New Year.
This thread has definitely deviated from its original intent. Now we're discussing vodka. Great.
Millwood, you really should have started a new thread with your MOSFET JLH thing, I'm sure you'd get the same participation from tschrama, and you wouldn't have messed up this one.
If you're really only looking for a use for your MOSFET devices, have you considered Nelson's FET version of the Citation 12?
http://www.diyaudio.com/forums/showthread.php?threadid=21866&highlight=
Edit: From a search, I see you are aware of this so assume you've already tried it. Adapting a BJT amp to use FETs seems to be something of a compulsion for you...maybe start a new thread on that, I'm sure there would be lots of interest.
Millwood, you really should have started a new thread with your MOSFET JLH thing, I'm sure you'd get the same participation from tschrama, and you wouldn't have messed up this one.
If you're really only looking for a use for your MOSFET devices, have you considered Nelson's FET version of the Citation 12?
http://www.diyaudio.com/forums/showthread.php?threadid=21866&highlight=
Edit: From a search, I see you are aware of this so assume you've already tried it. Adapting a BJT amp to use FETs seems to be something of a compulsion for you...maybe start a new thread on that, I'm sure there would be lots of interest.
I am the rudest!!
Sorry Millwood I guess I win for being the rudest. But how about we don't jump down each others' throats for assumed slights. You really do tend to OVER REACT that way. Let's just all get along in this new year okay.
Hi Paul. How was your holiday? I have been sick (and miserable) the whole time. Let's get together soon.
Greg
Sorry Millwood I guess I win for being the rudest. But how about we don't jump down each others' throats for assumed slights. You really do tend to OVER REACT that way. Let's just all get along in this new year okay.
Hi Paul. How was your holiday? I have been sick (and miserable) the whole time. Let's get together soon.
Greg
Hi Greg, I sent an email to your work; I don't have your home email and it's disabled on the board here.
I'm still planning on building a JLH but was hoping to see what Graham Maynard's improvements look like. I may play with some slightly different topologies before settling on one, but no MOSFETs, I'll build a Zen if I want a FET amp. I'm hoping to put my JLH in that Technics case that I posted here a few pages back.
I'm still planning on building a JLH but was hoping to see what Graham Maynard's improvements look like. I may play with some slightly different topologies before settling on one, but no MOSFETs, I'll build a Zen if I want a FET amp. I'm hoping to put my JLH in that Technics case that I posted here a few pages back.
did more comparison between my mosfet and bjt versions of jlh.
I put a 1uf film cap on the speaker termals to see how each performs. The bjt version sounded dry with the cap added: the base was gone and there is a faint high frequency "hissing" in the background.
the same 1uf didn't impact the mosfet at all. Also, while taking off the cap with the amp still powered on, I saw sparks flying between the cap wires and the terminal. So I used one wire of the cap to touch the terminal while the other wire is firmed attached to the terminal. I got tons of sparks, and no sound degradation.
Both sounded very good without the cap.
also played around with a complementory version of the mosfet (similar to the Class-AB JLH amp tschrama had posted, however, with just one driver stage). it sounded also very good. but seemed to lack a little bit of "fluid-likeness" of the jlh1969.
I put a 1uf film cap on the speaker termals to see how each performs. The bjt version sounded dry with the cap added: the base was gone and there is a faint high frequency "hissing" in the background.
the same 1uf didn't impact the mosfet at all. Also, while taking off the cap with the amp still powered on, I saw sparks flying between the cap wires and the terminal. So I used one wire of the cap to touch the terminal while the other wire is firmed attached to the terminal. I got tons of sparks, and no sound degradation.
Both sounded very good without the cap.
also played around with a complementory version of the mosfet (similar to the Class-AB JLH amp tschrama had posted, however, with just one driver stage). it sounded also very good. but seemed to lack a little bit of "fluid-likeness" of the jlh1969.
the williamson tube amp and voltage driven output in jlh
so we talked about how "wrong" it is to use voltage driven devices in the jlh, and how the JLh was not designed for such devices.
Well, the root and inspiration of the JLH lies in the williamson tube amp. How does the Willaimson amp look like? I attached a picture here for you.
As you can see, it is an all triode amp in the input and driver stage. Like the JLH, it has a phase splitter (V2) driven a push-pull output stage (V3 and V4). There is no reliance on "current sharing" between V3 and V4 for it to work as tubes are voltage driven devices (less so than mosfets actually).
essentially, JLH did a Williamson impersonation with BJTs (current driven devices).
one of the shortcomings of the Williamson amp is the asymmetric output impedence of the phase splitter, which was also inherent in the JLH design.
so we talked about how "wrong" it is to use voltage driven devices in the jlh, and how the JLh was not designed for such devices.
Well, the root and inspiration of the JLH lies in the williamson tube amp. How does the Willaimson amp look like? I attached a picture here for you.
As you can see, it is an all triode amp in the input and driver stage. Like the JLH, it has a phase splitter (V2) driven a push-pull output stage (V3 and V4). There is no reliance on "current sharing" between V3 and V4 for it to work as tubes are voltage driven devices (less so than mosfets actually).
essentially, JLH did a Williamson impersonation with BJTs (current driven devices).
one of the shortcomings of the Williamson amp is the asymmetric output impedence of the phase splitter, which was also inherent in the JLH design.
Attachments
well spotted
I often wondered where the idea came from. very interesting.
for me this whole discussion of mosfet, bjt and now valve, has cleared up a few points of the subtlties of how this cct works.
For that reason my vote is that we keep all these discussions in the same thread.
mike
I often wondered where the idea came from. very interesting.
for me this whole discussion of mosfet, bjt and now valve, has cleared up a few points of the subtlties of how this cct works.
For that reason my vote is that we keep all these discussions in the same thread.
mike
millwood said:
It has been done succesfully and reported to JLH in 1969 but beyond that I don't know any details
Geoff would almost certainly know more on this.
I tried a few transistors on my breadboarded jlh1969 and would like to share my view with you all.
first of all, the overall schematic is the same as the jlh1969, with a 2n5401 upfront, a mje15030 in the middle. the emitter resistor on the mje15030 is set at 150ohm, to run the driver stage a little bit hotter and to accomodate the mosfet output devices. bias current set around 750ma for the output devices. the source is a portable cd player line-out through a parasound preamp (php850).
the bjt output devices that I have tried:
mje15030, mjl21194, 3281 and 4281, vs. tip41c.
the mosfets are:
irfp240 and fdp3682 vs. irf540.
of the bjts, I liked mje15030 and 21194 the best. they are clear, vibrant and very lively. 3281 and 4281 are simply indistinguishable from the tip (not bad at all to remind you).
my favorite mosfet is still the 540. irfp240 sounded just as good but it is a pain to mount two of them on one small heatsink. fdp3682 doesn't sound as full and seems to be quite fragile: I blew two of them while setting the bias (trembling hand on the trimmer 🙂).
however, the differences were not huge and I could flip-flop on quite a few combinations. the only thing clear is that
a) with the tip in the driver, the sound was dull and a little muffled - however, the tip gives decent sound at the output.
b) 2sc4793 has got to be the brightest transistors I have ever heard.
I eventually settled for a mosfet driver (a fairchild medium power) and two 540s: they sound very good (not overly bright yet vibrant) and are much easier to mount on my skinny cpu heatsink than those bigger devices.
first of all, the overall schematic is the same as the jlh1969, with a 2n5401 upfront, a mje15030 in the middle. the emitter resistor on the mje15030 is set at 150ohm, to run the driver stage a little bit hotter and to accomodate the mosfet output devices. bias current set around 750ma for the output devices. the source is a portable cd player line-out through a parasound preamp (php850).
the bjt output devices that I have tried:
mje15030, mjl21194, 3281 and 4281, vs. tip41c.
the mosfets are:
irfp240 and fdp3682 vs. irf540.
of the bjts, I liked mje15030 and 21194 the best. they are clear, vibrant and very lively. 3281 and 4281 are simply indistinguishable from the tip (not bad at all to remind you).
my favorite mosfet is still the 540. irfp240 sounded just as good but it is a pain to mount two of them on one small heatsink. fdp3682 doesn't sound as full and seems to be quite fragile: I blew two of them while setting the bias (trembling hand on the trimmer 🙂).
however, the differences were not huge and I could flip-flop on quite a few combinations. the only thing clear is that
a) with the tip in the driver, the sound was dull and a little muffled - however, the tip gives decent sound at the output.
b) 2sc4793 has got to be the brightest transistors I have ever heard.
I eventually settled for a mosfet driver (a fairchild medium power) and two 540s: they sound very good (not overly bright yet vibrant) and are much easier to mount on my skinny cpu heatsink than those bigger devices.
assembly question
Hi All,
im still building the chassis and making Dutch DIYs boards. Thanks Dutch🙂
But I have read somewher in this huge thread that turn on/off thump may be present in this amp.
Would the following elektor circuit solve the turn on thump?
http://mitglied.lycos.de/Promitheus/delay_circuit_for_toroids.htm
If it does I cant see how it would fix turn off thump, is there any simple way to cure this?
any help appreciated🙂
Hi All,
im still building the chassis and making Dutch DIYs boards. Thanks Dutch🙂
But I have read somewher in this huge thread that turn on/off thump may be present in this amp.
Would the following elektor circuit solve the turn on thump?
http://mitglied.lycos.de/Promitheus/delay_circuit_for_toroids.htm
If it does I cant see how it would fix turn off thump, is there any simple way to cure this?
any help appreciated🙂
Re: assembly question
two "turn-on" thumps are talked about here:
1) turn-on current rush for the mosfet version: I have found no cure for it. The saving grace is that there is no "pop" in the speaker. so no audioable impact.
2) turn-on thump in the bjt version: some have ran into it, others haven't. I have not see it myself. The turn-on current rush is much smaller in the bjt version than in the mosfet version.
As to the source of turn-on thump, I think it is due to the input decoupling cap charging up. so it is likely that there will be no cure for it unless you raise the rail voltage slowly.
I would build the amp first and then the turn-on circuitry if you run into the turn-on thump.
Luke said:
two "turn-on" thumps are talked about here:
1) turn-on current rush for the mosfet version: I have found no cure for it. The saving grace is that there is no "pop" in the speaker. so no audioable impact.
2) turn-on thump in the bjt version: some have ran into it, others haven't. I have not see it myself. The turn-on current rush is much smaller in the bjt version than in the mosfet version.
As to the source of turn-on thump, I think it is due to the input decoupling cap charging up. so it is likely that there will be no cure for it unless you raise the rail voltage slowly.
I would build the amp first and then the turn-on circuitry if you run into the turn-on thump.