No, i didn't wonder. I know exactly what causing it and know how to remove it.
I think the GM is popular among people from your area. I remembered it was Guanghao who promoted the circuit. It shouldn't be difficult to find the schematics.
I have built the Mimesis, but i have also added extra driver for each output pair, just like Telos. I have never had peaking like yours, because of the extra compensation i mentioned are part of the original Mimesis. Telos should have it too i believe, because my knowledge tells me that you have to if you want to have good sound from this topology.
And that 80v for both front end and output stage (latfet) just didn't make sense for a good design.
Why did you add the extra driver? Is it only about spreading the heat and saving heat sinks, as mentioned a few posts above by keantoken?
I presumed the current was set by:
the resistor (from 220 to 100 Ohms),
the rail voltage (± 50 to 90),
the bias setting.
So in the DC condition, a second or third pair would each have to dissipate the same heat as the first pair.
If I correctly recall my measurements across the resistor (and if I measured correctly!), in the dynamic condition, the voltage drop across it remained the same, therefore the current remained the same and the drivers didn't go into class B.
The resistor was 220 Ohms, I don't recall the voltage drop across it. Rails were about ±70. Six lateral MOSFETs and one pair of drivers.
Does this Telios / Mimesis design work a bit differently from that?
I presumed the current was set by:
the resistor (from 220 to 100 Ohms),
the rail voltage (± 50 to 90),
the bias setting.
So in the DC condition, a second or third pair would each have to dissipate the same heat as the first pair.
If I correctly recall my measurements across the resistor (and if I measured correctly!), in the dynamic condition, the voltage drop across it remained the same, therefore the current remained the same and the drivers didn't go into class B.
The resistor was 220 Ohms, I don't recall the voltage drop across it. Rails were about ±70. Six lateral MOSFETs and one pair of drivers.
Does this Telios / Mimesis design work a bit differently from that?
To measure the Vdrop across the 220 Ohm, I simulated the dynamic condition by setting a DC voltage at the amplifier input, then measuring the voltage and various places, changing the input voltage and measuring again. The effect was a very low frequency changing voltage.
Also, just remembered, a friend told me he changed the 220 to a 180 Ohm and his amplifier sounded better, "fuller and warmer".
I presume that change would allow the gate capacitance to discharge through the resistor, and charge through it, a bit faster, or with less voltage drop as that happened?
With 4, 6 or 8 MOSFETs and one pair of drivers, there is more capacitative charge to move as the voltage swings, so might that increase a Vdrop across the resistor?
With more drivers, ie, 1 pair per pair of MOSFETs, that multiple of capacitance charge wouldn't happen. Would that make a better (more linear) performing stage and better sound quality?
Also, just remembered, a friend told me he changed the 220 to a 180 Ohm and his amplifier sounded better, "fuller and warmer".
I presume that change would allow the gate capacitance to discharge through the resistor, and charge through it, a bit faster, or with less voltage drop as that happened?
With 4, 6 or 8 MOSFETs and one pair of drivers, there is more capacitative charge to move as the voltage swings, so might that increase a Vdrop across the resistor?
With more drivers, ie, 1 pair per pair of MOSFETs, that multiple of capacitance charge wouldn't happen. Would that make a better (more linear) performing stage and better sound quality?
I would rather look for an alternative theory. The suggestion the designer is complicating the topology and pcb routing in order to save cents in a device of this quality and stature does not sound serious.
Hi can you decrease r13 , 330 ohm so high this is increase feedback and oscillation ,maybe must be 150 ohm @10ma
where do you found R24 R27 ? 100nF is on each driver
http://past.hifipig.com/wp-content/uploads/2015/06/20150212_161556-1.jpg
http://past.hifipig.com/wp-content/uploads/2015/06/20150212_161556-1.jpg
Last edited:
Are you saying that you did no other change at that time, just added the extra drivers, and it sounded better? 🙂
Did you keep the same value resistor? 220 Ohms or another size?
Last edited:
I have never built the real Goldmund (same with many other designs). I started with simulating the real Goldmund Mimesis 3 (I think the resistor is 330R not 220R), studying it, noting its performances, then did my own work from there.
I have the original JFET but never used it. I used a metal can jfet with gold plated pins used in McIntosh preamp.
The way i build amps is using simulation. Sometimes i didn't use the real models, just using trusted models with similar key parameters. Based on my understanding of how an amp works, i made change to the circuit, change the prototype and listen. My final decision is based on listening. The more i do this, the more i see relationship between sound i hear and what i see in simulation.
If you are curious about the 220R and how reducing it can 'improve' the sound, i will share a little experience... If you change the resistor, just like that without clear objective, there are many parameters changed in the circuit and you will never know which one is responsible with the change in sound. But from all of my amps, i tend to have minimal Vgs or Vbe (As you can see, reducing the R may reduce gate to gate voltage, or simply the Vgs). In my latest amp, based on simulation i found that to improve performance i need to change something. I made the change in the prototype, measured some operating points and listened. I could hear improvement but i was so curious because the 'Vbe' of the driver was only 500 something mV. I checked the datasheet and it showed that minimum Vbe is 700mV. But i kept it under 600mV because that what my ears wanted. What i'm telling you is, i'm relying on my ears and i tend to have minimal Vbe and Vgs in the output stage.
You will find them in PC_03 document. They are two 10R resistor as parallel per each side.
Each driver has a coupling cap 100nF. There are no output coil. Just available in newer models after (next gen and near ones).
Telos uses MJE340/350 as drivers. If you run only one pair to enough charge/discharge MOSFETs, it must be biased at high currents. Output sounding can be worse. Slight biasing sound more open.
After reading the first 25 pages of posts on this thread, before work today, I thought about it all day long. Audio, engineering artistry, porn. KUDOS to all contributing. It is so easy to see and understand the many facets of beauty that makes up the Goldmund, legacy / sound. Taking notes and studying circuit topologies intently...
Goldmund 27+ diy
Hi would this be any good ?
bought with a job lot on ebay .
i have removed 1 of the 4 2N5566 dual fets and it was very closely matched .
I am currently building a Mezmerize B1 buffer .
would it be worthwhile building this in to a chassis or just stick with the Mezmerize build .
20190621_145946 by glenn jarrett, on Flickr
20190621_150020 by glenn jarrett, on Flickr
20190619_141649 by glenn jarrett, on Flickr
Hi would this be any good ?
bought with a job lot on ebay .
i have removed 1 of the 4 2N5566 dual fets and it was very closely matched .
I am currently building a Mezmerize B1 buffer .
would it be worthwhile building this in to a chassis or just stick with the Mezmerize build .
20190621_145946 by glenn jarrett, on Flickr20190621_150020 by glenn jarrett, on Flickr20190619_141649 by glenn jarrett, on FlickrHi would this be any good ?
bought with a job lot on ebay .
i have removed 1 of the 4 2N5566 dual fets and it was very closely matched .
I am currently building a Mezmerize B1 buffer .
would it be worthwhile building this in to a chassis or just stick with the Mezmerize build .
Seems made in China? Where did you get the PCB?
How's the sound? 🙂
Looks like this may be the only source at the moment.
Imitation gaowen (GOLDMUND) 27 preamp amplifier PCB (empty board)
I thought I had seen a completed one at one time on ebay.
Imitation gaowen (GOLDMUND) 27 preamp amplifier PCB (empty board)
I thought I had seen a completed one at one time on ebay.
Has anyone facts/info about the original Goldmund (Job) SRA power amp? I am finding it difficult to get manual and schema. I would also appreciate a lot if someone could explain (to a non-engineer) what the so called job circuitry was/is all about?
I recently acquired a mint unit and rather like the sound.
I recently acquired a mint unit and rather like the sound.