Interesting....But for sure the output laterals must be closely matched to run without resistors....I know the reverse temperature coefficient helps smoothing the current distribution... but anyway. Closely matching must be of vital importance.
The resistor less design may aid the performance towards the notorious more grunt of the BJT's...
The resistor less design may aid the performance towards the notorious more grunt of the BJT's...
over the majority of the frequency range the -ve shunt has lower output impedance and this is almost certainly explained by the improved performance of the NPN Medium Power pass transistor.Salas;2331906 post56 said:
Why does the performance at the extreme LF end and the extreme HF ends go in favour of the +ve shunt?
Even more so with Goldmund, the founding father is not a designer/EE but a marketing guy.
(i believe Michel Reverchon was PR something at IBM's before he got a taste for gold)
What is amusing about the Goldmund prime believers thread is that the company is originally French, as is Michel Reverchon, not Swiss.
Goldmund's single endeavour for the first 8 years or so of existence was turntables.
The second fun part is that Mr Reverchon didn't do the TT development stuff himself, but hired French platine gurus for the job, the main one being Pierre Lurne of Audiomeca.
The company relocated to Switserland around 1987 for tax reasons and the "Swiss" aura.
Same time frame as the 1st Goldmund loudspeaker and amp products, the Mimesis 2+3.
You got to admire Mr Reverchon's marketing talent though.
What really shocks is the response from the DIY lot : just mention Gold/Swiss/Expensive, say it's the very best amp, and they're sold (lock, stock and barrel).
i have a buddy who is an engineer for another swiss audio and video company. he shared that the amp business is not making any money, its legacynbecause of the founders heritage and tradition but from a pure p&l perspective it is not a significant moneymaker. he is not with goldmund but with another high end swiss company.
his credentials are impeccable, he was one of the early broadcast engineers for bbc and knows audio and particularly since the seventies video inside out (a frequent speaker at conventions and standards comittee meetings). i guess what i want to say is that not every swiss company is built in the same way as goldmund.
im in the design and engineering of high tech, i know very well that price is often a strong indication of marketing funds than hard research dollar, especially for the companies that wont let you see their engineering labs or doesnt much of their own ip in terms of patents or research papers.
back to regular programming, sorry for the interruption...
his credentials are impeccable, he was one of the early broadcast engineers for bbc and knows audio and particularly since the seventies video inside out (a frequent speaker at conventions and standards comittee meetings). i guess what i want to say is that not every swiss company is built in the same way as goldmund.
im in the design and engineering of high tech, i know very well that price is often a strong indication of marketing funds than hard research dollar, especially for the companies that wont let you see their engineering labs or doesnt much of their own ip in terms of patents or research papers.
back to regular programming, sorry for the interruption...
Last edited:
The message was that it's hardly a Swiss enterprise, and nowadays it's Monaco based.
SBM Offshore has a headquarter in Switserland, operations are in Monaco as far back as some of my former class mates did their thesis research there.
Still is a prominently Dutch business, though the management can probably only say Cheese.
The hottest item for iTunes now is the LaRosita Black Hole, hand made and sold by Christian Yvon (also French and designer of the Apologue LSP)
A $27.5K fun item, Mr Yvon will likely have to sell a few to get a refund on his development expense.
Not quite the same.
High bandwith/slew hitachi mod.
Since nobody else has offered a direction , may I intervene ?? Inspired by mr. Chua at Ampslab.com and his flagship Bi240 ,what I got was a hitachi based topology that has 120+ V/us slew (everbody's requirement) ,a OLG of over 72DB (lee's issue) and is stable with 87 degrees margin at a UG of 890KHZ (pretty high bandwidth). All I had to do is alter the VAS with EF's (below) and degenerate the stages appropriately.
Still tweaking it , but it has possibilities. THD 20 .0015% / 1K - under .0008% @ 40v p/p. About twice as good as my original symasym (CX) , with over 20db more gain.
OS
Since nobody else has offered a direction , may I intervene ?? Inspired by mr. Chua at Ampslab.com and his flagship Bi240 ,what I got was a hitachi based topology that has 120+ V/us slew (everbody's requirement) ,a OLG of over 72DB (lee's issue) and is stable with 87 degrees margin at a UG of 890KHZ (pretty high bandwidth). All I had to do is alter the VAS with EF's (below) and degenerate the stages appropriately.
Still tweaking it , but it has possibilities. THD 20 .0015% / 1K - under .0008% @ 40v p/p. About twice as good as my original symasym (CX) , with over 20db more gain.
OS
Attachments
Target Specifications?
From simulations it would appear to be good for around 200 watt into 8 ohms resistive.
The THD is not stunning but simulates at about 0.02% at 22V rms into 8 ohm. Predominantly 2nd harmonic, which is maybe why it is claimed to sound good. Maybe there are other things contributing to it being good.
I am not crazy about the fact that phase change is 180 degrees at 1.4 MHz while the gain remains about 13 dB, hence the concern that some members raised regarding stability.
I also would like to point out that there is no input decoupling which may lead to the output off-set to be 31 x the input off-set, meaning that your source has to be zero off-set.
Slew rate is about 136 V/uS using 10uS period square wave, again simulated whether this is practical is debatable. It should be measured in a real amp that includes all parasitic and lumped constants contained in the PCB layout.
Running an output noise simulation there is a definite peak in response at 418kHz which can be assumed as oscillation, so this circuit does border on instability even with a resistive load. Be aware that the ferrite beads on the MOSFET legs may be an absolute necessity. Besides there are beads and beads, pick what is required for the application.
Output noise simulates to be 300nV/root Hz, again it is a simulation. So is the weather and it is hardly something to go by.
One should accept simulations as indicative only under absolutely perfect conditions as there are no lead lengths, track to track and component to component interaction, power supply noise, earthing problems, all components at the same temperature or other nonconformities in the simulation regardless of how accurate the models are. Besides simulations assumes that every component with the same credentials are identical, which is hardly practical.
Whether these are the targeted specifications is really hard to tell, and probably not. But is the only indication of where the paper design is.
Perceived sound quality is totally subjective and has little to do with the simulated characteristics. A half deaf person may think the amp performs impeccably, while his pets and all the birds in the garden flees when hearing it. The opposite can also apply.
On the other hand, taking care and cloning the product exactly, at least will closely match that what is being sold and what some perceive as the best amplifier they ever heard. Assuming of course that everyone shares the same perception, listen to the same music, in the same envoronment, using the same speakers, and will come to the same conclusion.
Kind regards
Nico
From simulations it would appear to be good for around 200 watt into 8 ohms resistive.
The THD is not stunning but simulates at about 0.02% at 22V rms into 8 ohm. Predominantly 2nd harmonic, which is maybe why it is claimed to sound good. Maybe there are other things contributing to it being good.
I am not crazy about the fact that phase change is 180 degrees at 1.4 MHz while the gain remains about 13 dB, hence the concern that some members raised regarding stability.
I also would like to point out that there is no input decoupling which may lead to the output off-set to be 31 x the input off-set, meaning that your source has to be zero off-set.
Slew rate is about 136 V/uS using 10uS period square wave, again simulated whether this is practical is debatable. It should be measured in a real amp that includes all parasitic and lumped constants contained in the PCB layout.
Running an output noise simulation there is a definite peak in response at 418kHz which can be assumed as oscillation, so this circuit does border on instability even with a resistive load. Be aware that the ferrite beads on the MOSFET legs may be an absolute necessity. Besides there are beads and beads, pick what is required for the application.
Output noise simulates to be 300nV/root Hz, again it is a simulation. So is the weather and it is hardly something to go by.
One should accept simulations as indicative only under absolutely perfect conditions as there are no lead lengths, track to track and component to component interaction, power supply noise, earthing problems, all components at the same temperature or other nonconformities in the simulation regardless of how accurate the models are. Besides simulations assumes that every component with the same credentials are identical, which is hardly practical.
Whether these are the targeted specifications is really hard to tell, and probably not. But is the only indication of where the paper design is.
Perceived sound quality is totally subjective and has little to do with the simulated characteristics. A half deaf person may think the amp performs impeccably, while his pets and all the birds in the garden flees when hearing it. The opposite can also apply.
On the other hand, taking care and cloning the product exactly, at least will closely match that what is being sold and what some perceive as the best amplifier they ever heard. Assuming of course that everyone shares the same perception, listen to the same music, in the same envoronment, using the same speakers, and will come to the same conclusion.
Kind regards
Nico
The above was simulating the Goldmund as it stands in the Goldmund thread. I would like to suggest that one modification is done at a time and discussed instead of a whole host of mods in one shot, that way everyone can keep up with pace at which things change.
Maybe we should also decide what is the most important parameters that needs to be achieved. It was summed up very well in the first post and we should put them in order of impotance and solve them as we progress.
Regards
Nico
Maybe we should also decide what is the most important parameters that needs to be achieved. It was summed up very well in the first post and we should put them in order of impotance and solve them as we progress.
Regards
Nico
Last edited:
J201 fairchild low noise Jfet (40v Vgs) , cheap at mouser , says "made for audio"
Product Folder - J201 - N-Channel General Purpose Amplifier, Fairchild Semiconductor - Global Leader in Power Optimization
At the "high end" 110v p/p (77vrms- 392w , truly fictional), I get .008% @1k and worst .012% THD20. About 4 X as good as my big symasym.
I already am on my way to replacing the CX with this , at least in prototype form (CX1.2/3) so , I can try it on my 200w BJT OPS.
CCS adjust stepping (PIX1) 2.25 mA is the best at the "tails"
(winner - pix2/.0006%)
LT below "CX_1.2fet.zip"
Edit : this would be a "no brainer" to port to driving laterals ... eassssy!!
OS
Attachments
Last edited:
This thread is hilarious and pointless!!! You guys do realize that this is NOT improving the Goldmund amp, but creating something else entirely? Right? If something new no longer resembles the original, then what is the purpose of this discussion?
May I kindly suggest:
Why don't you all agree on the specs, the cost, the topology (MOSFET/Bi Polar), etc. that you'd like the amplifier to have and then just build the circuit that will give those specs?
That's how real engineers do it
May I kindly suggest:
Why don't you all agree on the specs, the cost, the topology (MOSFET/Bi Polar), etc. that you'd like the amplifier to have and then just build the circuit that will give those specs?
That's how real engineers do it
No offence, but isn't it enough that this thread is separate?
We know you like the original design. We also know that the original design is flawed and this thread is about correcting those flaws. Maybe it will be a totally different amp, maybe not. Just let us do our thing OK?
May I kindly suggest:
Why don't you all agree on the specs, the cost, the topology (MOSFET/Bi Polar), etc. that you'd like the amplifier to have and then just build the circuit that will give those specs?
That's how real engineers do it
I actually agree with NagysAudio.
But I am still interested to see where this design will go because I have a soft heart for this topology.
I will be building both versions to see what difference I might hear. As well as compare to my version of this topology as well as JLH's version and Trace Elliots version.
Last edited:
REAL Engineers are doing right here what REAL engineers do
1. conceive
2. see what has been done before
3. integrate LEARNED improvements
4. design up
5. test
6. change design to improve the results
7. test
go back to step 3 and continue until it is as close to perfection as time permits
THEN
Sit back and enjoy the fruits of their labor, thoughts and time well spent and laugh at the amateurs that try to imitate the REAL ENGINEERS
1. conceive
2. see what has been done before
3. integrate LEARNED improvements
4. design up
5. test
6. change design to improve the results
7. test
go back to step 3 and continue until it is as close to perfection as time permits
THEN
Sit back and enjoy the fruits of their labor, thoughts and time well spent and laugh at the amateurs that try to imitate the REAL ENGINEERS