Multiple input ampilfiers.
Xavier,
I have seen some articles about using multiple parrallel input stages each using multiple input devices. The outcome was an increased noise performance of 2.7dB for 2 parallel stages and 5.1 dB for 4 parallel stages.
The output of all stages was summed by the output stage.
Would that be an option for you to increase the amplification?
Xavier,
I have seen some articles about using multiple parrallel input stages each using multiple input devices. The outcome was an increased noise performance of 2.7dB for 2 parallel stages and 5.1 dB for 4 parallel stages.
The output of all stages was summed by the output stage.
Would that be an option for you to increase the amplification?
Thanks Promitheus and Tarasque for your replies. Right now, these are the 4 possibilities I am studying (please correct me if I'm wrong):
1- Step in transformer
Although a good option, most of the opinions I read from people that tested both the transformer and a MC active pre tend to prefer the latest (tighter bass, clearer sound, etc.).
2- Additional Input stage before the existing input
See http://www.diyaudio.com/forums/showthread.php?s=&threadid=24834&highlight=.
I have some concerns on the distortion and noise performance. The output of this headamp would be more than 4 mV and this would go to the input of the Pearl. According to the fig.16 in Wayne's article (distortion curve), this input would produce a distortion of 0,6%, that could easily become 1% or more depending of the input voltage and the frequency.
3- Additional jFets in parallel
This is the option that Tarasque was pointing. The Pearl "as-is" already has 4 parallel devices. We could duplicate that to a total of 8 parallel devices. In theory this would be the best option since we still have only 2 gain stages. The noise performance should be better as well.
But, aren't we going to have problems because of the differences between the jFets, the more devices, the more dificult to match them? Is there a limit on the number of jfets to be added in parallel?
What would be the performance in terms of distortion for both the first gain stage and the second stage after the RIAA section?
4- Additional output stage after the existing output.
A single stage using a IRF610 mos (based in the Aleph preamp). This would keep the distortion very low. My concern would be more in terms of noise that this stage could generate, but I might work on the filtering of the PSU to minimize that.
So, to answer Tarasque's question, yes 3- would be an option, but does anyone know if 8 parallel devices is too much? and what would be the expected distortion after increasing the gain for the first stage and increasing then the input level for the second stage?
BTW I have 2 extra matched quads 2sk170 to test this, that Promitheus sent to me. But unfortunately I don't think the quads are matched between them to generate 2 sets of 8 matched devices!
Xavier
1- Step in transformer
Although a good option, most of the opinions I read from people that tested both the transformer and a MC active pre tend to prefer the latest (tighter bass, clearer sound, etc.).
2- Additional Input stage before the existing input
See http://www.diyaudio.com/forums/showthread.php?s=&threadid=24834&highlight=.
I have some concerns on the distortion and noise performance. The output of this headamp would be more than 4 mV and this would go to the input of the Pearl. According to the fig.16 in Wayne's article (distortion curve), this input would produce a distortion of 0,6%, that could easily become 1% or more depending of the input voltage and the frequency.
3- Additional jFets in parallel
This is the option that Tarasque was pointing. The Pearl "as-is" already has 4 parallel devices. We could duplicate that to a total of 8 parallel devices. In theory this would be the best option since we still have only 2 gain stages. The noise performance should be better as well.
But, aren't we going to have problems because of the differences between the jFets, the more devices, the more dificult to match them? Is there a limit on the number of jfets to be added in parallel?
What would be the performance in terms of distortion for both the first gain stage and the second stage after the RIAA section?
4- Additional output stage after the existing output.
A single stage using a IRF610 mos (based in the Aleph preamp). This would keep the distortion very low. My concern would be more in terms of noise that this stage could generate, but I might work on the filtering of the PSU to minimize that.
So, to answer Tarasque's question, yes 3- would be an option, but does anyone know if 8 parallel devices is too much? and what would be the expected distortion after increasing the gain for the first stage and increasing then the input level for the second stage?
BTW I have 2 extra matched quads 2sk170 to test this, that Promitheus sent to me. But unfortunately I don't think the quads are matched between them to generate 2 sets of 8 matched devices!
Xavier
So, to answer Tarasque's question, yes 3- would be an option
Xavier,
The study I mentioned is summing the signal after the cascode but before the output stage.
You can still use the quad input fets and as the output of the cascode is again showing uncorrelated noise, adding parralel input stages and summing after the cascode would again create nose cancellation.
Xavier,
The study I mentioned is summing the signal after the cascode but before the output stage.
You can still use the quad input fets and as the output of the cascode is again showing uncorrelated noise, adding parralel input stages and summing after the cascode would again create nose cancellation.
Tarasque,
I think now I got what you were saying. You are not only paralleling the jfet devices but the whole stage including the cascode.
Let's assume we have two stages in parallel having each one a gain of 30 dB. What would be the resulting gain of this?
Can I get more information on this study?
Thanks
I think now I got what you were saying. You are not only paralleling the jfet devices but the whole stage including the cascode.
Let's assume we have two stages in parallel having each one a gain of 30 dB. What would be the resulting gain of this?
Can I get more information on this study?
Thanks
Xavier,
Two amplifiers with the same gain summed will result in 6dB voltage gain.
Have a look at Jourmal of audio eng socity vol 35, no11, 1987
D. Self / Ultra-low-noise amplifiers and granularity distortion
Selfs result:
Gain +69dB, Noise -81dBu, EIN-150dBu with 4 parallel amplifiers and 2 parallel gain stages per amp.
Two amplifiers with the same gain summed will result in 6dB voltage gain.
Have a look at Jourmal of audio eng socity vol 35, no11, 1987
D. Self / Ultra-low-noise amplifiers and granularity distortion
Selfs result:
Gain +69dB, Noise -81dBu, EIN-150dBu with 4 parallel amplifiers and 2 parallel gain stages per amp.
Raising the Gain of the Pearl for MC Cartridge's...
Raising the gain in any way, will in theory raise the noise floor as well and S/N Ratio will be lowered, but isnt' it so on all MC phono preamps compared to the same preamp as an MM phono preamp.
promitheus has an really good aproach by means of an MC Transformer, but an really good one is usually very expencive
Paralleling Stages as posted by Tarasque, might work ?, but is, in my opinion, a expencive solution as well.
Additional jFets in parallel. This could be a suitable solution, but as pointed out, which problems are we facing other than the hard time matching more than 4 jfet's, ohh my what a pool we need ( for 16 matched transistors (2 x 8)).
Further more we then have to think about the jFet's caracteristics if we match more than the 4 pr channel.
Perhaps just my humble opinion, but have given this issue alot of thaught, and i think the best solution might be to build the ONO's MC stage, and ad it to the Pearl's input, with some sort of selector for MM and MC.
analogair
Raising the gain in any way, will in theory raise the noise floor as well and S/N Ratio will be lowered, but isnt' it so on all MC phono preamps compared to the same preamp as an MM phono preamp.
promitheus has an really good aproach by means of an MC Transformer, but an really good one is usually very expencive
Paralleling Stages as posted by Tarasque, might work ?, but is, in my opinion, a expencive solution as well.
Additional jFets in parallel. This could be a suitable solution, but as pointed out, which problems are we facing other than the hard time matching more than 4 jfet's, ohh my what a pool we need ( for 16 matched transistors (2 x 8)).
Further more we then have to think about the jFet's caracteristics if we match more than the 4 pr channel.
Perhaps just my humble opinion, but have given this issue alot of thaught, and i think the best solution might be to build the ONO's MC stage, and ad it to the Pearl's input, with some sort of selector for MM and MC.
analogair
I dont think a high end aproach like using a Lundahl Tranformer is so expensive.
it only costs 100 Euro for 2 pieces and you dont have any problems.
Or just use a better MM catridge.
If you invest the whole work and the costs of making a better MC stage just use a better MM.
An expensive MM can play better than a cheap MC.
it only costs 100 Euro for 2 pieces and you dont have any problems.
Or just use a better MM catridge.
If you invest the whole work and the costs of making a better MC stage just use a better MM.
An expensive MM can play better than a cheap MC.
Raising the Gain of the Pearl for MC Cartridge's...
Yes, but we must also take in consideration the fun of building the darn thing by our self promitheus.
I would like to build the MC stage of the ONO, and add it to the Pearl. My only prob. is that i'm not familiar with doing PCB layout.
Anyone Anyone up for it ?
http://www.passlabs.com/downloads/old product/aonoserv0.pdf
Page 5 of 9
analogair
Yes, but we must also take in consideration the fun of building the darn thing by our self promitheus.
I would like to build the MC stage of the ONO, and add it to the Pearl. My only prob. is that i'm not familiar with doing PCB layout.
Anyone Anyone up for it ?
http://www.passlabs.com/downloads/old product/aonoserv0.pdf
Page 5 of 9
analogair
I'm using the Pearl and the Lundahl....
transformers now along with the Denon DL-103R. Its an excellent combination, IMHO.
While I've thought about an active stage of gain much like Wayne discusses in the Pearl documentation, I've not been too motivated to make any changes to the Pearl's boards.
Currently, I have step-up transformers in a separate case with their own loading resistors so that I can remove it from the audio chain quickly when I switch pre-amps around or swap cartridges.
I actually prefer this Pearl/Denon pairing with the Pearl over the tube phono preamp sections in my McIntosh MX-110 or the Mac C2200 preamp.
Cheers,
David
transformers now along with the Denon DL-103R. Its an excellent combination, IMHO.
While I've thought about an active stage of gain much like Wayne discusses in the Pearl documentation, I've not been too motivated to make any changes to the Pearl's boards.
Currently, I have step-up transformers in a separate case with their own loading resistors so that I can remove it from the audio chain quickly when I switch pre-amps around or swap cartridges.
I actually prefer this Pearl/Denon pairing with the Pearl over the tube phono preamp sections in my McIntosh MX-110 or the Mac C2200 preamp.
Cheers,
David
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.