Hey Martin,
Kevin mentioned D3a. Splendid choice that is frequently on sale at ebay.de.
The russian 6C45 is also easy to find there.
Another interesting tube is the triode-pentode 6F12P. This one has a triode with mu=100and Gm=19mA, the pentode when trioded has the same steepness and mu=70. Have a bunch of them at the bench that I have measured in my AVO. Still haven´t tried it in any circuits though.
If you to go by the ECC- tubes, the 88 is my choice.
Kevin mentioned D3a. Splendid choice that is frequently on sale at ebay.de.
The russian 6C45 is also easy to find there.
Another interesting tube is the triode-pentode 6F12P. This one has a triode with mu=100and Gm=19mA, the pentode when trioded has the same steepness and mu=70. Have a bunch of them at the bench that I have measured in my AVO. Still haven´t tried it in any circuits though.
If you to go by the ECC- tubes, the 88 is my choice.
Thanks for the suggestions.
@Sy
yes, I know 'FragJanZuerst' - a good recommendation, thanks. It is a very good shop, he has great transformers with reasonable prices and the F&T MKP caps he sells are a good choice to my ears. I just did not find a 6GK5 there (maybe I should look a bit more carefully).
If I understand you correctly, you would dump the entire RCA topology and go for something completely different.
I followed some of your projects here with interest and trust your recommendations. Any hint which way to go (other than the RCA topology)?
Keep in mind that I would like to have a +/- simple topology.
I am still in the process of re-reading Morgan Jones' book.
@revintage
would you use the D3A triode-connected or as pentode?
Thanks,
Martin
@Sy
yes, I know 'FragJanZuerst' - a good recommendation, thanks. It is a very good shop, he has great transformers with reasonable prices and the F&T MKP caps he sells are a good choice to my ears. I just did not find a 6GK5 there (maybe I should look a bit more carefully).
If I understand you correctly, you would dump the entire RCA topology and go for something completely different.
I followed some of your projects here with interest and trust your recommendations. Any hint which way to go (other than the RCA topology)?
Keep in mind that I would like to have a +/- simple topology.
I am still in the process of re-reading Morgan Jones' book.
@revintage
would you use the D3A triode-connected or as pentode?
Thanks,
Martin
Don't think you'll find them on mainland Europe. Mostly in the States and Japan. Therefore you might as well go for the D3a. A German tube. (Triode-connected)
Martin,
The D3a could be used either way but Zout will be higher going pentode. Think Dave Slagle did some calculations and found Ra=12k (=Zout) being the best. In this case probably triode-strapping is the best.
The D3a could be used either way but Zout will be higher going pentode. Think Dave Slagle did some calculations and found Ra=12k (=Zout) being the best. In this case probably triode-strapping is the best.
Martin, it's not so much the topology as the implementation- a passive RIAA sandwiched between two gain stages is very viable, just not the RCA version (designed for cheep'n'easy). I have a phono preamp project which will be appearing shortly. I would also recommend any of Allen Wright's designs, Kevin Kennedy's phono stage (some great similarities to mine), or Gregg the Geek's. If you have Valve Amplifiers, the phono stage designs in there are excellent, and use ECC88 or 5842 as input tubes.
@ Eli
Salas makes a very good point about the location of that first coupling capacitor although I just tend to make it pretty large (and expensive.. 🙄 ) To minimize interaction with rp over the life of the tube you do want to make R1 about 10X larger than rp, but don't get carried away with this because the johnson noise of R1 brings up the LF noise floor significantly. (More of an issue truthfully with very quiet high transconductance tubes because you don't want to degrade the noise performance with a resistor that is much noisier than the tube driving it..)
My take on the Lipshitz circuit is a little odd, I place the bias resistor before the actual network right after the coupling cap - because this eliminates the issue with dividing down the gain as a ratio of R1/Rgrid which because I don't use grid leak bias is a significant effect.
You need to take the thevenin equivalent of rp+R1//20M in your calculations for the equalizer network - though its effect will be small. Whether or not you do as Salas suggests make sure the pole introduced by the coupling capacitor and R is <5Hz for good results.
I have seen Sy's latest design and would strongly recommend it as being well worth the wait. SY and I have generally similar approaches to this problem and I believe his design will be a little easier to get working properly. My current design has not been published to date, (it's almost 2 yrs old now) but will probably eventually show up on Positive Feedback if I can get my act together and write an article about it. 😛
The D3A in triode mode is a spectacular performer. It really redefines the meaning of the word quiet. The johnson noise due to winding resistance of a typical MM cartridge will probably be 10 - 12dB noisier than this tube, a good high output MC (low dcr) cartridge might be just a few dB quieter than the tube. The typical 12AX7A based front end is typically at least 3dB - 6dB noisier than the noisiest MM types IME.
Not to steal my own thunder, but.. agreed on the MM, disagree on the MC. Input referred equivalent noise resistance of the triode-connected D3a is 65-100R. Now a low DCR cartridge will be something like 3R, right? That works out to about 13-14dB more noise from the tube compared to the cartridge, if I'm doing the math correctly. And that doesn't count the noise from the mandatory grid-stopper.
Hi SY,
I was thinking and writing with regard to high output MC which typically have some tens of ohms of dcr, not the low output types which are generally much lower. The old Sumiko Blue Point as an example has a dcr of around 30 ohms if memory serves. I have one around here but cannot find it to check. (Old and broken cantilever) Some current Sumiko HO MC types are listed by the manufacturer as having this high a level of dcr, but information is not always easy to find. Contrast this with some MM which have a dcr of >600 ohms or more. Note the key here is I wrote "High Output MC" - these usually have much higher dcr than a low or medium output MC. (Which IMO would need a transformer despite acceptable noise performance, due to insufficient gain at least in my design.) Also I should have clarified low dcr relative to typical MM types which is what I meant.
Here is a listing of Benz MC types with dcr included - spans the range from 5 ohms to 120 ohms and includes quite a few LO/MO/HO types in that range: http://www.musicalsurroundings.com/cartr.html
Incidentally it does appear that there are some HO MC types that will be noisier than the tube. (dcr in the vicinity of 120 ohms) Still very quiet overall.
Last edited:
Hi all
Speaking of phonostage input tubes, anyone tried/has an opinion about EC/PC88 or 86?
They can be had for free (almost) and seems spec-wise quite OK.
/Olof
Speaking of phonostage input tubes, anyone tried/has an opinion about EC/PC88 or 86?
They can be had for free (almost) and seems spec-wise quite OK.
/Olof
Kevin, agreed. The ones I had in mind were at the low end of the range. My own cartridge has the unhappy combination of low output (0.2mV) and high-ish resistance (15R). That means even with a perfect preamp, the s/n is limited to 69dB or so.
Hi SY,
I'm assuming that there are probably a couple of other noise sources in the vinyl playback chain that are running neck in neck with your cartridge. Surface/tracing noise and rumble right on the disk you are playing come to mind, plus local rumble and other noises. I suspect in the real world there is probably less than a 10dB margin between your cartridge and the best that can be achieved.
I suspect if I worked the math on any of those benz micro I cited there are only a couple that do significantly better than yours. (5 ohms 0.2mV output models at $5K come to mind - ouch... )
I suspect some of the HO MC types in the range of 2mV out and 30 ohms would be near the theoretical nadir.
Transformers have their own issues and I suspect the secondary dcr in all but the very best transformers (probably permalloy) that step up 20dB or so could dominate over the cartridge noise source.
None of this probably matters much as I have very few disks that even outperform the thermal noise floor of my Grado cartridge, and when you factor in the complementary effect of the RIAA eq on the noise floor the playback system is quieter than any of my records..
PC88 datasheet sates an Req of about 240R which is among the best, common to the double triode 6922 family. It brings you down to 2nV/rtHz. If you parallel 2, you go to about 1.42nV/rtHz. Comparable to a single low noise Jfet. A C3M trioded gets down to 2.6nV/rtHz for comparison. Better do that with a PCC88 though which has the single envelope advantage, keeping things tight and common Tc.
Actually the transformers will be at their best with medium impedance carts for 47k spec load secondaries when fed into low noise triode or fet inputs. For new good pressed records expect 65dB SNR with 70dB as tops if direct metal mastered.
Not to beat the D3A drum again, but I'm gonna 😛 the internal equivalent noise source of the triode connected D3A is about 60 - 70 ohms which puts it in the 1nV/rtHz region which is extremely quiet. It will be quieter than most phono cartridges including a lot of MC types. Were I really smart I would keep my mouth shut and lay in a life time supply..
I agree, that's in line with my experience.. Not very impressive when we quote dnr and noise floor numbers for good digital playback systems.. 😀 However I will still take good vinyl playback over CD any day, 24/96 and SACD is a much harder sell.
OT
I'm actually listening to a downloaded (legally) Czech Radio 44.1K/16bit flac recording of a live Czech state orchestra performance of the Brandenburg Concertos and the sound is very, very good. Minimally processed, surprisingly it sounds great. (Rather good performance too)
Last edited:
Noise is a nuisance, not the whole story...😉
They have their Velvet celebrations, (nothing to do with Lou Reed).
They have their Velvet celebrations, (nothing to do with Lou Reed).
Keep your mouth shut.😀
While you're at it, you might mention the astonishing linearity at high voltage swings. The first time I had one on the spectrum analyzer, I thought my equipment was broken.
- Status
- Not open for further replies.