In the 70s I first used MIALs, and later Siemens ps and pp. The Siemens were a decided improvement.
John, I don't know if you feel like Scott offended you, but it seems like you are laying it on kind of thick. Too much of that risks backfiring, and risks making you look unreasonably grouchy. Probably better all around if we can work towards keeping the tone as civil as possible.
Last edited:
I believe Scott's numbers my sims come very close to what he measures. Looks like a great DIY project. Having done some very simple DIY phonos the simpler the better. I like the caps over the servo one high quality part. Nothing against servos but they do take parts and builders come in all skill levels from beginner to way smarter than I am.
Scott seemed to be saying that all the phono preamp circuits come from the past, including the basis for his. That's the way I read it, anyway.
PS are very fine. also large and expensive and often hard to find. But fine. Cheaper and every bit as good is the dc servo. small and cheap... just the way you like it.
THx-RNMarsh
Last edited:
I think you are right, Markw4. I did misunderstand Scott's input. Personally, I think that everybody should work together on a project like this. Of course, not ALL phono circuits come from the past, there are rather advanced modern ones as well. I personally tend not to release my latest designs for discussion, for example.
EUVL, your question is noted. However, the STAX approach is direct coupled, with push pull vacuum tube drive for the headphones, and it does not require additional capacitors to isolate the headphone bias. This is pretty darn optimum.
Now, could it be better? Yes, better parts, better DC power supply, better connectors, etc, etc. I know this would make a difference, but it is pretty good as it is. I already usually bypass the Alps pot at the input. (I can hear it) Yet, and I have heard even better, newer, more expensive versions of the STAX Headphone elements on occasion, but I just can't afford the upgrade.
Richard, it's true that servos are cheap, cheap, cheap! Compared to high quality polystyrenes. '-)
EUVL, your question is noted. However, the STAX approach is direct coupled, with push pull vacuum tube drive for the headphones, and it does not require additional capacitors to isolate the headphone bias. This is pretty darn optimum.
Now, could it be better? Yes, better parts, better DC power supply, better connectors, etc, etc. I know this would make a difference, but it is pretty good as it is. I already usually bypass the Alps pot at the input. (I can hear it) Yet, and I have heard even better, newer, more expensive versions of the STAX Headphone elements on occasion, but I just can't afford the upgrade.
Richard, it's true that servos are cheap, cheap, cheap! Compared to high quality polystyrenes. '-)
if the circuit has potential for wide variations in DC offset when built and/or drifts a lot, then coupling cap might be better way to go. A good bipolar when topology allows.
PSty caps btw have the lowest dielectric absorption as w.jung and I first pointed out long ago and very stable. But now the values desired are harder to find, larger than other types and more expensive. But for RIAA they are the best choice.
A good push-pull tube power amp with 400V on plates can be used with stax hdp. Just cap couple off the plates to each side of electro panel grid.. I did this for years on my tri-amped system for the RTR ESL tweeters. decades ago. Then you just need a HV dc power supply to bias the center element. No transformers to go thru.
-RNM
PSty caps btw have the lowest dielectric absorption as w.jung and I first pointed out long ago and very stable. But now the values desired are harder to find, larger than other types and more expensive. But for RIAA they are the best choice.
A good push-pull tube power amp with 400V on plates can be used with stax hdp. Just cap couple off the plates to each side of electro panel grid.. I did this for years on my tri-amped system for the RTR ESL tweeters. decades ago. Then you just need a HV dc power supply to bias the center element. No transformers to go thru.
-RNM
Last edited:
Yes, I was just repeating that there is nothing new here just a simplification down to as little as you need. Over the last few years Jan has published lots of phono articles some quite complex, for differential phono this is about as simplistic as Salas' long running thread.
As for the servo issue I don't have a single sound card or digital recorder that isn't AC coupled so in fact I could just remove the caps anyway.
I still have the question how do you mix single and dual supply devices without coupling caps? Professional microphones are usually 48V single supply phantom power they all use coupling caps so like op-amps they're already in the chain. Does the average servoed PA take 12V DC from a 24V single supply line amp?
Last edited:
47uf at 63V that's what the pros use. I asked, how do you servo out 48V phantom power it's at the very front of virtually every modern recording chain? Just curious would make a splash I would think.
http://www.thatcorp.com/datashts/dn140.pdf
With film caps there's still microphonics I have both a mic preamp and the QA400 that are actually bothersome for microphonics. Actually tapping on my keyboard 3' away shows up in the noise floor if either is on the bench.
http://www.thatcorp.com/datashts/dn140.pdf
With film caps there's still microphonics I have both a mic preamp and the QA400 that are actually bothersome for microphonics. Actually tapping on my keyboard 3' away shows up in the noise floor if either is on the bench.
Last edited:
Thx for posting the schematic, Scott!
I am one of the ones who made a cascoded folded cascode cascoded (Scott has seen it), all BJT, and I also added another 3 pairs of JFETs to the IPS as to get lower noise as my version is for 64dB gain at 1kHz.
Having 84dB gain in one stage in my experience causes DC offset trouble when temperature varies in the room, so I am trying to figure out how to add DC servos to make both output have zero offset (no coupling caps).
I also added source resistors hoping to get lower dist.
THD1k is <100 dB (0,8Vrms output voltage SE) in sim but will surely be a bit worse IRL. Often not worse than 10dB, though, in my experience.
But I did not build it as first I want a version that can deal with DC offset properly...
I am one of the ones who made a cascoded folded cascode cascoded (Scott has seen it), all BJT, and I also added another 3 pairs of JFETs to the IPS as to get lower noise as my version is for 64dB gain at 1kHz.
Having 84dB gain in one stage in my experience causes DC offset trouble when temperature varies in the room, so I am trying to figure out how to add DC servos to make both output have zero offset (no coupling caps).
I also added source resistors hoping to get lower dist.
THD1k is <100 dB (0,8Vrms output voltage SE) in sim but will surely be a bit worse IRL. Often not worse than 10dB, though, in my experience.
But I did not build it as first I want a version that can deal with DC offset properly...
Well, at least we have people talking about quality jfet based phono stages!
That is certainly an improvement.
I personally think that the 2 gain block phono circuit works even better than any single gain stage, due to signal handling concerns, so I will stick to it.
Complexity? You guys have no idea about complexity! Design elegance is first, then optimization of each gain stage (cascode included). There are measurable advantages to optimizing each gain stage for current, complexity, buffering, etc. I do find that just for a static measurement like 1KHz thd, some go too complex and actually make the circuit measure better but subjectively sound worse by adding additional junctions that the audio signal has to traverse.
At this level of audio quality, why not just use quality IC's?
That is certainly an improvement.
I personally think that the 2 gain block phono circuit works even better than any single gain stage, due to signal handling concerns, so I will stick to it.
Complexity? You guys have no idea about complexity! Design elegance is first, then optimization of each gain stage (cascode included). There are measurable advantages to optimizing each gain stage for current, complexity, buffering, etc. I do find that just for a static measurement like 1KHz thd, some go too complex and actually make the circuit measure better but subjectively sound worse by adding additional junctions that the audio signal has to traverse.
At this level of audio quality, why not just use quality IC's?
A quick comment to those that were interested in the differential phono I posted recently. I populated two boards and assembled them in an enclosure fully connectorized. This could be the most transparent phono stage that I have had, I have played several "difficult" LP's with no issues.
A note on the gain. I started out making something to put in front of an external sound card or field recorder with low noise mic preamps built in. There was a provision on my card, not on the schematic, for a low noise instrumentation amplifier to convert the diff-out to single ended and that's what I auditioned. A THAT 1512 at a gain of 5 (14dB) would take you to 40dB at 1kHz and have virtually no noise penalty.
EUVL has sent me some suggestions for different FET's on the input and a reworking of component values to get 40dB or so out of the original topology. These seem fine and I would have no issue in trying it this way. And of course for MC there have to be changes. Personally I think the benefit of the differential interface is most important on the input side so I had no issue with a conventional IC as a second stage. At these high impedance levels the FET buffers eliminate the current noise of the IC in-amp as an issue.
Anyone has my blessing to do anything they want with this. it was an experiment in how far simplicity can get you (Zen was taken so I said Muntz). An open-loop diff-pair and a totally shunt de-emphasis, I have not seen a lot of those so it was a chance to try something very different. We probably don't want to carry on here in any case.
A note on the gain. I started out making something to put in front of an external sound card or field recorder with low noise mic preamps built in. There was a provision on my card, not on the schematic, for a low noise instrumentation amplifier to convert the diff-out to single ended and that's what I auditioned. A THAT 1512 at a gain of 5 (14dB) would take you to 40dB at 1kHz and have virtually no noise penalty.
EUVL has sent me some suggestions for different FET's on the input and a reworking of component values to get 40dB or so out of the original topology. These seem fine and I would have no issue in trying it this way. And of course for MC there have to be changes. Personally I think the benefit of the differential interface is most important on the input side so I had no issue with a conventional IC as a second stage. At these high impedance levels the FET buffers eliminate the current noise of the IC in-amp as an issue.
Anyone has my blessing to do anything they want with this. it was an experiment in how far simplicity can get you (Zen was taken so I said Muntz
). An open-loop diff-pair and a totally shunt de-emphasis, I have not seen a lot of those so it was a chance to try something very different. We probably don't want to carry on here in any case.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.