John,
I appreciate I could learn a bit of your design philosophy and experience. I believe I have absorbed a little piece of it and would be able to follow and continue.
I appreciate I could learn a bit of your design philosophy and experience. I believe I have absorbed a little piece of it and would be able to follow and continue.
That's because you have a mid-fi phono playback system, Scott. I like FM too, but I also know its limitations.
With lower quality system, one usually appreciates signal source with less resolution and dynamic. That is one of the reasons why FM is sometimes considered to be a good sound source. Sound smoother than higher resolution source on worse system.
PMA,
I lost track somewhere along the line when you were printing graphs without schematics. As of your most recent postings, are the caps across the folded cascode reference resistors still in place?
Being a hopeless troglodyte in the eyes of some folks, I didn't simulate these things, I just tried them. Lower resistances mean lower noise and caps short that noise to ground and stabilize the DC to the MOSFET Gates. Obviously, this all goes out the window if you hook up the folded cascode the way John did in the JC-3, using both the Gate/base and Source/emitter. On the other hand, you can stick some pretty low resistance values into the first stage load position, so you can lower the noise that way, too.
One of the things I'm looking forward to in the next iteration of the GR-25 is being able to lower the loads for the front end because I will no longer have to account for two to four volts worth of Vgs in order to get the MOSFETs to bias. Yes, the front end gain will drop, but who gives a rip? I'm dumping gain right and left anyway and the vast majority of it comes from the MOSFETs/bipolars in the folded cascode actively loading each other.
I have finally begun serious fiddling with the discrete phono stage I've been messing with off and on for the past couple of years. The gain and noise aspects will be more of an issue for me than with the amp.
Incidentally, won't increasing the value of the capacitances for the folded cascode lower the corner frequency for the noise shelf?
Grey
I lost track somewhere along the line when you were printing graphs without schematics. As of your most recent postings, are the caps across the folded cascode reference resistors still in place?
Being a hopeless troglodyte in the eyes of some folks, I didn't simulate these things, I just tried them. Lower resistances mean lower noise and caps short that noise to ground and stabilize the DC to the MOSFET Gates. Obviously, this all goes out the window if you hook up the folded cascode the way John did in the JC-3, using both the Gate/base and Source/emitter. On the other hand, you can stick some pretty low resistance values into the first stage load position, so you can lower the noise that way, too.
One of the things I'm looking forward to in the next iteration of the GR-25 is being able to lower the loads for the front end because I will no longer have to account for two to four volts worth of Vgs in order to get the MOSFETs to bias. Yes, the front end gain will drop, but who gives a rip? I'm dumping gain right and left anyway and the vast majority of it comes from the MOSFETs/bipolars in the folded cascode actively loading each other.
I have finally begun serious fiddling with the discrete phono stage I've been messing with off and on for the past couple of years. The gain and noise aspects will be more of an issue for me than with the amp.
Incidentally, won't increasing the value of the capacitances for the folded cascode lower the corner frequency for the noise shelf?
Grey
Grey, please don't think so much. It is getting confusing. First of all, the JC-3 circuit is not really a folded cascode because it is dominated by the common source connection.
At this rate, without thinking it through, you can make serious errors.
At this rate, without thinking it through, you can make serious errors.
I had asked you once what you wanted to call the circuit, as "that thing kinda like a folded cascode that John uses" is somewhat clumsy as a moniker. I don't recall that you ever responded, but it would help to have a more concise name.
I agree that the input at the Gate dominates, particularly in cases where the feed to the Source comes from a resistor of only 10 Ohms or so, but on the other hand, it's not as though the connection isn't there, so there is something of the folded cascode involved. In my previous post I don't think I made clear that I was speaking of a "John's cascode" in the second paragraph. I don't have the schematic at hand, but I seem to recall that the JC-3 presented something like a 220 Ohm resistor to the base of the cascode and 10 Ohms to the emitter. Granted, 220 Ohms isn't all that unusual a load for a first stage, but it's better (noise-wise) than something closer to 1k which is what I'm faced with using MOSFETs, just to get them to bias. Going to bipolars will let me bring that front end load down somewhat. It'll go from the gee-it-would-be-nice category to the omigod-it's-crucial category in the phono stage. I've narrowed it down to four topologies I want to try there, but anything that creates noise in a phono stage is going to be a bother. Gain, however, will most definitely be welcome in that application.
Grey
I agree that the input at the Gate dominates, particularly in cases where the feed to the Source comes from a resistor of only 10 Ohms or so, but on the other hand, it's not as though the connection isn't there, so there is something of the folded cascode involved. In my previous post I don't think I made clear that I was speaking of a "John's cascode" in the second paragraph. I don't have the schematic at hand, but I seem to recall that the JC-3 presented something like a 220 Ohm resistor to the base of the cascode and 10 Ohms to the emitter. Granted, 220 Ohms isn't all that unusual a load for a first stage, but it's better (noise-wise) than something closer to 1k which is what I'm faced with using MOSFETs, just to get them to bias. Going to bipolars will let me bring that front end load down somewhat. It'll go from the gee-it-would-be-nice category to the omigod-it's-crucial category in the phono stage. I've narrowed it down to four topologies I want to try there, but anything that creates noise in a phono stage is going to be a bother. Gain, however, will most definitely be welcome in that application.
Grey
Grey,
parallel capacitor divider took place only in one simulation, in my reply to you, i.e. here:
http://www.diyaudio.com/forums/showthread.php?postid=1493527#post1493527
Following simulations were without caps, according to John request and directions.
parallel capacitor divider took place only in one simulation, in my reply to you, i.e. here:
http://www.diyaudio.com/forums/showthread.php?postid=1493527#post1493527
Following simulations were without caps, according to John request and directions.
Grey, actually a larger value resistor can be quieter, all else being equal, because more first stage gain is developed. See what I mean? It is important to learn the subject thoroughly through immersion into practical noise theory and its tradeoffs. Sometimes it is not as obvious as it first might seem.
Hi,
After trying for 20 minutes to filter the 170+ pages of this thread, I figured I would just ask.
Was the schematic, board layout, and/or parts list ever published, or is it still a secret?
Thanks
After trying for 20 minutes to filter the 170+ pages of this thread, I figured I would just ask.
Was the schematic, board layout, and/or parts list ever published, or is it still a secret?
Thanks
No, I did NOT publish the exact schematic. It is my intellectual property. However, the essence of the design has been derived, both by visual inspection of the circuit boards, and some good guesses.
We are now going forward with discussion of more intrinsic design principles that are usually more interesting to experienced engineers in the audio field.
We are now going forward with discussion of more intrinsic design principles that are usually more interesting to experienced engineers in the audio field.
PMA said:
Thanks. It's not a problem either way...I'm just trying to match your simulations to my thought experiments/experience. I do a lot of this by taking things to extremes and seeing what happens--somewhat like your 50000uF cap a couple of pages back as you attempted to get rid of phase shift. I do this while driving to work or on the way home because that's the quietest time I get all day, and no, I haven't caused any accidents...yet.
It helps that everyone scatters when they see me coming.
john curl said:
Again, agreed...assuming that I'm following you properly (honestly, last night was a booger at work and I was only able to give this matter a fraction of the attention it deserves--worse yet, I can hardly remember my train of thought at the time--rats), this is an avenue I've already been exploring for my discrete phono stage. I'm following this 75% from a perspective of noise and 25% from a gut feeling that I'd rather let the front end JFETs dominate in terms of distortion components, rather than the cascode. Percentages subject to change with further experiments.
The idea of a unity gain first stage (not all that uncommon in amps) showed me that way was folly...you add noise (and distortion, for that matter), but get nothing in return. Well, obviously that's stupid. After that it was a hop, skip, and a jump to the realization that there was going to be a optimum load value that gave positive gain, submerged the noise, and provided a decent Zout leading into the next stage.
While it's nice to have preamps and amps be as quiet as possible, it's absolutely essential for phono stages to be silent. Got any hints, tips, rules of thumb, suggested (available) books, or work-arounds that you'd care to share?
Grey
Grey, phono first stages are really tough. The most important thing is to not use mosfets in the gain path, or even the power supply buffer. Mosfets are just too inherently noisy, to keep them from adding noise, especially low frequency noise. However, in many cases, jfets can be used for 2'd stages, without too much trouble.
john curl said:
Thank you John, Actually I include a system of a Goldmund Reference table, $7000 Koetsu cartridge, and the best tube pre available at the time in that comparison. As well as John Dennison's demo of the JC-80, Lynn, Dennon 103(?) those two systems almost matched the master tapes over FM.
John,
The MOSFETs were just for the GR-25 and there only because I wanted the current. I tried JFETs in the cascode position in the amp but was unable to get a reasonable amount of current so the MOSFETs went in. The phono stage is all JFET at present, although I can see a couple of places where I might consider low noise bipolars. Maybe.
Grey
The MOSFETs were just for the GR-25 and there only because I wanted the current. I tried JFETs in the cascode position in the amp but was unable to get a reasonable amount of current so the MOSFETs went in. The phono stage is all JFET at present, although I can see a couple of places where I might consider low noise bipolars. Maybe.
Grey
john curl said:
Of course not, I got to listen to both at length though I have to admit the source material on the second was poorly chosen.
As I have pointed out on numerous occasions many very hightly regarded cartridges (last I checked the Denon DL-103 still has a spherical stylus) are physically incapable of playing many recordings without HUGE amounts of distortion. This is easily demonsrtrated with a 1/3 octave noise band on any good test record, the IM distortion sprays everywhere. I have only owned a Monster Alpha One (years ago it died) and various Grado References. I like the Grado's because you can get down to a small fraction of the coil's thermal noise pretty easily. They all 'crackle' on a few KHz and up 1/3 octave band. As they say on the Telarc Omni Disk "this is normal".
john curl said:
Pick one, any one, I picked the Denon because there are some people here who seem to have taste and an opinion worth respecting who still like this cartridge. I have never owned one and I find all the fiddling involved with a totally new cartridge to be just tedious. The Grados always have 10Hz resonance in my arm and behave the same as 35yr ago.
- Status
- Not open for further replies.