There comes a time now and then when I drastically change something about my stereo, or my computer, my house, my life etc.
This time it came in the form of a new phono line stage. I'm sick of the ones that come inside integrated amps and receivers and decided to try something completely different.
The idea was to get the line stage on its own and design it specifically for my own purposes.
To start with; a little about my setup: I'm using an original Pioneer PL600 turntable with a new Grado Prestige Black cartridge on it. (If you search for PL600 on google, you tend to come up with a much worse looking table, but anyway...)
To reduce the pickup of noise between the cartridge and the phono amp, I've changed everything to balanced operation. The cart operates un-grounded, with shielded twisted pair right out to the balun transformers.
For transformers I'm using Hammond 850G isolation transformers. My choice of transformers instead of active balanced inputs is that of common mode rejection. The nature of a high quality transformer is simply lightyears ahead in this area. Also, operating the transformers at such low level allows exceedingly low distortion on their part, and excellent frequency response.
The first stage is very simply an emitter follower to match the impedance of the cartridge and transformer to the impedance of the secend stage. The follower is loaded by a resistance, but will later be loaded with a current sink for greater accuracy and more linear loading.
The second stage is, I suppose, a little unusual as far as typical phono stages go. I've matched sets of four NPN transistors to go in parallel as a low noise class A gainstage. They are loaded by a very high impedance current source, aided by an opamp and transistor based servo; thus allowing very high gain and very low distortion. The transistors are biased with inherent DC local feedback via the collector load. It's really more or less a typical common emitter stage, with some embellishments to allow good performance.
The third stage is opamp based, I know.. yuck. It'll be discrete later.
It is based on a very ordinary opamp, the LM324 quad chip. It operates at very low gain with the RIAA filter in its feedback loop to correct the transfer function to be flat. After this stage is direct output to whatever front end I decide to load it with.
I will be removing the LM324 and building a discrete operational amplifier in its place to achieve yet even lower noise.
Perhaps when I get my scanner set up, I'll even post a schematic of what I've done. But those with engineering minds should be able to get a good idea from my above explanation.
The results? I have taken no measurements on this circuit as of yet. I'm more worried about the sound than the numbers. It does sound very good, and the noise floor is very very low. It's far lower than that of the typical integrated amps, even the high end ones. I now am able to clearly listen to things like tape print-through, turntable bearing noise, and intricate low-level details without any noticeable 'hiss' coming from the amplifier system at all. I'd say that's pretty good S/N ratio myself.
This time it came in the form of a new phono line stage. I'm sick of the ones that come inside integrated amps and receivers and decided to try something completely different.
The idea was to get the line stage on its own and design it specifically for my own purposes.
To start with; a little about my setup: I'm using an original Pioneer PL600 turntable with a new Grado Prestige Black cartridge on it. (If you search for PL600 on google, you tend to come up with a much worse looking table, but anyway...)
To reduce the pickup of noise between the cartridge and the phono amp, I've changed everything to balanced operation. The cart operates un-grounded, with shielded twisted pair right out to the balun transformers.
For transformers I'm using Hammond 850G isolation transformers. My choice of transformers instead of active balanced inputs is that of common mode rejection. The nature of a high quality transformer is simply lightyears ahead in this area. Also, operating the transformers at such low level allows exceedingly low distortion on their part, and excellent frequency response.
The first stage is very simply an emitter follower to match the impedance of the cartridge and transformer to the impedance of the secend stage. The follower is loaded by a resistance, but will later be loaded with a current sink for greater accuracy and more linear loading.
The second stage is, I suppose, a little unusual as far as typical phono stages go. I've matched sets of four NPN transistors to go in parallel as a low noise class A gainstage. They are loaded by a very high impedance current source, aided by an opamp and transistor based servo; thus allowing very high gain and very low distortion. The transistors are biased with inherent DC local feedback via the collector load. It's really more or less a typical common emitter stage, with some embellishments to allow good performance.
The third stage is opamp based, I know.. yuck. It'll be discrete later.
It is based on a very ordinary opamp, the LM324 quad chip. It operates at very low gain with the RIAA filter in its feedback loop to correct the transfer function to be flat. After this stage is direct output to whatever front end I decide to load it with.
I will be removing the LM324 and building a discrete operational amplifier in its place to achieve yet even lower noise.
Perhaps when I get my scanner set up, I'll even post a schematic of what I've done. But those with engineering minds should be able to get a good idea from my above explanation.
The results? I have taken no measurements on this circuit as of yet. I'm more worried about the sound than the numbers. It does sound very good, and the noise floor is very very low. It's far lower than that of the typical integrated amps, even the high end ones. I now am able to clearly listen to things like tape print-through, turntable bearing noise, and intricate low-level details without any noticeable 'hiss' coming from the amplifier system at all. I'd say that's pretty good S/N ratio myself.
How does an emitter follower input stage contribute to low noise?
What is it's purpose? Is the transformer providing no gain?
What is it's purpose? Is the transformer providing no gain?
As I said, the emitter follower is there to match impedance. The impedance of the cart through the transformer is 47K. The impedance of the common emitter amp is much lower.
If I wanted to get rid of the follower, I would have to have a step down transformer so that the impedance would not be so hard on the cartridge.
The purpose of the transformer is simply for isolation and to act as a balun (balanced to unbalanced).
This allows the longish lines between the phono cartridge and the amp to be balanced and thus pick up much less noise.
If I wanted to get rid of the follower, I would have to have a step down transformer so that the impedance would not be so hard on the cartridge.
The purpose of the transformer is simply for isolation and to act as a balun (balanced to unbalanced).
This allows the longish lines between the phono cartridge and the amp to be balanced and thus pick up much less noise.
Didn't realise it's an MM cart. I guess at MM levels one can do almost anything and still get away with acceptable noise.
Still, it's the wackiest phono topology i've ever hear of.
Still, it's the wackiest phono topology i've ever hear of.
That's actually been done in at least one successful commercial tube preamp- first stage is a follower to isolate the cartridge (or MC step up transformer) from the high input capacitance of a 12AX7.
analog_sa said:
Nothig wrong with wacky. It's not quite the way I'd do it, but who wants everything the same?
I have reason to believe that this preamp would still achieve good performance with a MC cartridge.
Later today I could easily to testing to determine it's S/N with that level of input.
Later today I could easily to testing to determine it's S/N with that level of input.
SY said:
SY
This is really interesting. I would imagine the first stage noise contribution is the most important and the less than unity gain certainly doesn't help. Which preamp is that?
I recall a patent on a similar phono preamp. I believe they used an emitter follower then a passive high frequency pole followed by a descrete amplifier with the feedback network providing the low frequency pole and mid frequency zero. I'll see if it's in my files.
Regards,
Ray
Regards,
Ray
Very interesting setup. That's sort of the next thing I was thinking of trying was to re-arrange where I set the poles in the amplifier.
This will happen when I get around to making the last stage discrete anyway.
This will happen when I get around to making the last stage discrete anyway.
When you've finished with the op-amp section you could always replace the phono stage in the car.
Just wondering, did you choose LM324 as a design choice or was it something that you've had in your junk drawer? This opamp is extremely unsuitable for audio because of the 7 kHz power bandwidth and extremely high crossover distortion.Duo said:
Lol, no I absolutely did not pick the 324 based on requirement.
It was just there on my bench. I don't like that opamp much at all.
It was just there on my bench. I don't like that opamp much at all.
... so what is your design goal then? Something that works, is good or extremely good? Something conventional or something odd or unusual?
- Status
- Not open for further replies.