What I'm talking about is not a simple RC stage..... but an active circuit. I have tried chokes but not on this sort of project. Since a simple pass transistor can make a cap look at least 10 times bigger(the multiplication of the caps value is based on the hfe of the device being used) you'd need a whole lot of choke to do what a 100,000 mfd cap could do. Change that simple pass transistor to a mosfet and it looks 10 times larger than the simple pass transistor version. Run that through another pass transistors and all thats left is a few microvolts of noise! Less space, weight, and a whole lot less cost! Take something like the power supply from the ARC PH-3 phono preamp. That power supply elevated to 300 volts would be totally suited to this sort of line stage. Cost to build the active circuit would be less than 25.00. This is why all the expensive iron just doesn't make sense to me....... I can achieve cleaner DC with cheap active circuitry and it would cost hundreds in iron and the results still wouldn't be as good.
Mark
Actually all comparisons were done using my old Gene Autry and Cab Calloway 78's that have been transfered to CD.....
. The Cabster sounds simply uncanny through the 12B4! The background noise from the 78's sounds extremely detailed and realistic..... Might as well can all this 12B4 stuff and just play the dang 78's!
Mark
. The Cabster sounds simply uncanny through the 12B4! The background noise from the 78's sounds extremely detailed and realistic..... Might as well can all this 12B4 stuff and just play the dang 78's!Mark
Same circuit in CD player output circuit
I'm late to this thread, but wanted to add that I have been using an almost identical 12B4 circuit as the output stage in a Pioneer DVD/CD player (voltage output DAC). I use 8.2K plate loads, but a CCS in the cathode (well bypassed) to hold current constant. Elaborate power supply filtering (LCRCRCLC). Sounds simply fabulous. No noise, even with headphones. I left the opamps in so I can switch back and forth to **** off solid-state guys. I'm about to do a similar treatment to my Sony SACD player. Screw the warranty!
I'm late to this thread, but wanted to add that I have been using an almost identical 12B4 circuit as the output stage in a Pioneer DVD/CD player (voltage output DAC). I use 8.2K plate loads, but a CCS in the cathode (well bypassed) to hold current constant. Elaborate power supply filtering (LCRCRCLC). Sounds simply fabulous. No noise, even with headphones. I left the opamps in so I can switch back and forth to **** off solid-state guys. I'm about to do a similar treatment to my Sony SACD player. Screw the warranty!
I've had my eye on the 7233 for line amp applications. I was thinking of trying an Akido-type setup, but using a FET-based current source load, as I have some Supertex HV TO-220 depletion-mode MOSFETs looking for something to do. Iwas also looking at using a triode strapped 6AQ5/6005, which would give me a bit more gain. I was looking at the 12B4 as well, but I only have 1-2 used units on hand of unkown condition right now, and I'd like to try out what I've got...
Well, In another thread about the GG preamp there was a comment from a mamber that the 12B4 is alot less expensive to build. I would completely disagree with that. It's at least as expensive when one takes into acocunt the power supply required for the 12B4.
Brian,
Would you share your CCS ideas with us?
Thanks!
Mark
That definitely wasn't me, as I've never posted on the GG preamp, and most common cathode stages have inherently poor PSRR when resistively loaded. A good quiet PS can be expensive of $ (unless you've been collecting iron for years as I have) and space, even if all it does is serve a "simple" circuit like the 12B4A one you posted. If you look at my previous posts on the 12B4 preamp I've been posting on here for at least 3 years, you'll see the lengths I go to to get the PS right, even in the predominantly balanced/differential circuits I prefer.Mark A. Gulbrandsen said:
Oh, nothing special at all. I frequently use a CCS in cathode circuits. I just use a good old LM317 and a current setting resistor, well bypassed with a big electrolytic and a small “boutique” film cap. The reason I bother with a CCS in the cathode is to hold the tube’s operating point constant over life and the same between channels. This centers the operating point of the plate voltage and keeps the plate resistance (output resistance) and transconductance close to the same for reasonably well-matched tubes. An LM317 costs maybe $0.49, and I have a bucket of them, so why not use it in many cases? No one should say that they can be heard in such a cathode circuit; the bypass caps completely hide any deleterious effects of a solid-state CCS. All the CCS has to do is to be a good current source at low frequencies, and any high frequency irregularities are hidden by the bypass caps which “short out” the CCS at all audio frequencies. I often use two separate such current sources in push-pull output stages to force primary current balance and ensure DC magnetization cancellation. No adjustments over tube life.
To complete the output filtering for the CD player, I used a passive LRC low-pass between the DAC and 12B4 grid to set a -3dB point at 58KHz, with only -0.026 dB impact at 20KHz.
A CCS as a plate load is a different beast and harder to get right, since you CAN hear the solid-state edge unless great care is taken (a la Gary Pimm’s designs). But great potential exists here, too.
No, just divide 1.25 volts by the desired current to define the current-setting resistor value. I see no reason why the current couldn't go as high as 1.5 amps, although this resistor gets quite small in ohmic value. I've used them happily up to 50mA in tube cathodes.
Edit: Oops - Just re-read you post and you did say "minimum" current, not "maximum". Yes, you are correct; there is a minimum current, that is often stated as 10mA, but in fact can be less if you carefully study the LM317 data sheets for the temperatures and voltages likely to be encountered. So it wouldn't be used to bias a 12AX7, but I tend to pull a lot of current through tubes like a 6DJ8 or 12B4 to get the transconductance up, so this limitation is not a problem for me generally. For my CD output stage application with a 12B4, I run almost 30mA through each tube to achieve the "text book" 1K plate resistance.
Edit: Oops - Just re-read you post and you did say "minimum" current, not "maximum". Yes, you are correct; there is a minimum current, that is often stated as 10mA, but in fact can be less if you carefully study the LM317 data sheets for the temperatures and voltages likely to be encountered. So it wouldn't be used to bias a 12AX7, but I tend to pull a lot of current through tubes like a 6DJ8 or 12B4 to get the transconductance up, so this limitation is not a problem for me generally. For my CD output stage application with a 12B4, I run almost 30mA through each tube to achieve the "text book" 1K plate resistance.
Mark A. Gulbrandsen said:
I'm glad that you find your preffered line stage
I'm completely (spell spelll spellll??) aware of arguments and metodology argumented in yours (above) lines,and I also tried all that stuff in my tube beginings ( senturies ago
) what is my preffered type of supply and type of line stage you can see if you click at www button on bottom of my post....
each approach have own arguments ,but only builder's taste is finally responsible for finall decision
hehe-to each his own
cheers!
Hi,
I was encouraged by our forum members especially Brett (Thanks Brett) and built a 12B4A line stage. It sounded wonderful. I initially used some lower quality parts. How it sounded almost completely depended on how good the filter caps I used. This was what I initially had:
380-0-380V Tx => Sovtek 5AR4 => 10H => 470uF (H.Q.) => (split into 2 channels) R => 330uF (L.Q. cap). I built a LM317 regulated heater supply. I didn't buy any caps and just used what I had on hands.
The sound was initially messy and hummed quit a bit. I found out that the 330uF caps were not good at all and replaced them with the Panasonic ED 100uF. The sound became pretty good but the hum was still there. Replaced the 100uF with ICW Clarity Cap (Polypropylene) 2 x 20uF = 40uF on each channel the HF sound was very satisfying but the bass was severely lacking. It hummed quite loudly and I couldn't enjoy the music. I then paralleled the 100uF with one ICW MKP 20uF for each channel and started enjoying the music since. It still had a bit of hum. The latest change was to use the 330uF low quality cap as the first cap, like this: 10H => 300uF (L.Q.) => R => 470uF (H.Q.) (split to 2 channels) => R => 100uF (H.Q.) + 20uF (MKP). What was the sound? still a touch of messiness, still a bit of hum, but overall - Excellent. I most certainly believe that sound wise it beats the Audio Research xxxx preamp my friend has that I liked so much which triggered me to look into building a tube preamp.
Then one of the tubes (Industrial grade Westinghouse) started making more and more noise (like firewood burning sort of cracking sound). Is that microphonic? leakage? I am pretty sure it was the tube because I swapped the tubes and the noise only came in the channel with the bad tube. It got so bad so I decided to shut it down and rebuild it.
I have ordered 4 x GE JAN Mil Spec 12B4A and 2 x Sylvania Mil Spec 12B4A and they will arrive at any time from now. I will compare their sound. I have bought 2 x 200uF Solen MKP for the last cap of the power supply. I am getting a new custom made Toroidal Tx to replace the old E.I. transformer as it hummed and gots VERY HOT! I am also replacing the 5AR4 because I don't like the preamp to run so hot with 5V x 2A = 10W wasted in rectification! I decided to make it mono stereo and have ordered 8 x UF4007 and the 8 x 0.01uF 1000V ceramic caps and will make 2 snubbers for the secondary wirings. 2 x 10H hammond chokes will arrive tomorrow. All filter capacitors are Panasonic TSHA and TSHB except the last caps being Solen MKP. It will be like this (mono): Tx 350V => Snubber => UF4007 => 10H => 100uF => 270R => 220uF => 270R => 470uF => 270R => 200uF (Solen). It will give me 300VDC running on 25mA + 2mA (breeder) with the AC ripples reduced to 0.02mA.
I have some questions about the layout of the preamp. I have one 2mm thick steel sheet to shield the Tx and chokes from the circuit in the steel chassis. How effective is such a shield? Can it eliminate (or reduce to an acceptable level of) RFI? If it can shield the RFI from the Tx and chokes, the energy must go somewhere. I guess it would be turned into currents in the chassis. How to do deal with it? I think a star ground should help.
Another question is whether I should make another 2 steel sheets and install them between the Tx and the 1st choke, and between the 1st choke and the 2nd choke. I may not have sufficient space to orient the chokes 90deg from each other. What is the best way to do it?
This is the new layout:
Thanks and regards,
Bill
I was encouraged by our forum members especially Brett (Thanks Brett) and built a 12B4A line stage. It sounded wonderful. I initially used some lower quality parts. How it sounded almost completely depended on how good the filter caps I used. This was what I initially had:
380-0-380V Tx => Sovtek 5AR4 => 10H => 470uF (H.Q.) => (split into 2 channels) R => 330uF (L.Q. cap). I built a LM317 regulated heater supply. I didn't buy any caps and just used what I had on hands.
The sound was initially messy and hummed quit a bit. I found out that the 330uF caps were not good at all and replaced them with the Panasonic ED 100uF. The sound became pretty good but the hum was still there. Replaced the 100uF with ICW Clarity Cap (Polypropylene) 2 x 20uF = 40uF on each channel the HF sound was very satisfying but the bass was severely lacking. It hummed quite loudly and I couldn't enjoy the music. I then paralleled the 100uF with one ICW MKP 20uF for each channel and started enjoying the music since. It still had a bit of hum. The latest change was to use the 330uF low quality cap as the first cap, like this: 10H => 300uF (L.Q.) => R => 470uF (H.Q.) (split to 2 channels) => R => 100uF (H.Q.) + 20uF (MKP). What was the sound? still a touch of messiness, still a bit of hum, but overall - Excellent. I most certainly believe that sound wise it beats the Audio Research xxxx preamp my friend has that I liked so much which triggered me to look into building a tube preamp.
Then one of the tubes (Industrial grade Westinghouse) started making more and more noise (like firewood burning sort of cracking sound). Is that microphonic? leakage? I am pretty sure it was the tube because I swapped the tubes and the noise only came in the channel with the bad tube. It got so bad so I decided to shut it down and rebuild it.
I have ordered 4 x GE JAN Mil Spec 12B4A and 2 x Sylvania Mil Spec 12B4A and they will arrive at any time from now. I will compare their sound. I have bought 2 x 200uF Solen MKP for the last cap of the power supply. I am getting a new custom made Toroidal Tx to replace the old E.I. transformer as it hummed and gots VERY HOT! I am also replacing the 5AR4 because I don't like the preamp to run so hot with 5V x 2A = 10W wasted in rectification! I decided to make it mono stereo and have ordered 8 x UF4007 and the 8 x 0.01uF 1000V ceramic caps and will make 2 snubbers for the secondary wirings. 2 x 10H hammond chokes will arrive tomorrow. All filter capacitors are Panasonic TSHA and TSHB except the last caps being Solen MKP. It will be like this (mono): Tx 350V => Snubber => UF4007 => 10H => 100uF => 270R => 220uF => 270R => 470uF => 270R => 200uF (Solen). It will give me 300VDC running on 25mA + 2mA (breeder) with the AC ripples reduced to 0.02mA.
I have some questions about the layout of the preamp. I have one 2mm thick steel sheet to shield the Tx and chokes from the circuit in the steel chassis. How effective is such a shield? Can it eliminate (or reduce to an acceptable level of) RFI? If it can shield the RFI from the Tx and chokes, the energy must go somewhere. I guess it would be turned into currents in the chassis. How to do deal with it? I think a star ground should help.
Another question is whether I should make another 2 steel sheets and install them between the Tx and the 1st choke, and between the 1st choke and the 2nd choke. I may not have sufficient space to orient the chokes 90deg from each other. What is the best way to do it?
This is the new layout:
An externally hosted image should be here but it was not working when we last tested it.
Thanks and regards,
Bill
My two cents about the 12b4 linestage:
I built one following the schematic on diyparadise.com; using 12k anode resistor and 260V B+. Power supply is dual mono, capacitor input double pi filter with European Rgn1064 mesh anode rectifiers (electrically comparable to 5y3 but the mesh version is a big step upwards from anything else) - Sound is excellent; quite close to what Yeo describes: Prominent high tones; amazing neutrality + gain in soundstage; especially the increased depth struck me. Power supply capacitors are all motor-run caps (pio) which are in every amp a big improvement on electrolytics. Power supply is unregulated and I plan to keep it that way in the definitive version as there aren't too many arguments (soundwise) for a regulated ps. The 0,33µF coupling caps on the schematic are big-time overkill; I work with 0,056 Russian military teflon-in-oil capacitors (still around 7hz cut-off so you won't miss anything unless your speakers are huge tml's or big subwoofers + you gain everything by using smaller, faster and better caps). I'm planning to play with the loadlines soon; altering B+ and anode resistor & maybe cathode resistor. Will keep you posted if I find any audible improvements on the current version.
Simon
I built one following the schematic on diyparadise.com; using 12k anode resistor and 260V B+. Power supply is dual mono, capacitor input double pi filter with European Rgn1064 mesh anode rectifiers (electrically comparable to 5y3 but the mesh version is a big step upwards from anything else) - Sound is excellent; quite close to what Yeo describes: Prominent high tones; amazing neutrality + gain in soundstage; especially the increased depth struck me. Power supply capacitors are all motor-run caps (pio) which are in every amp a big improvement on electrolytics. Power supply is unregulated and I plan to keep it that way in the definitive version as there aren't too many arguments (soundwise) for a regulated ps. The 0,33µF coupling caps on the schematic are big-time overkill; I work with 0,056 Russian military teflon-in-oil capacitors (still around 7hz cut-off so you won't miss anything unless your speakers are huge tml's or big subwoofers + you gain everything by using smaller, faster and better caps). I'm planning to play with the loadlines soon; altering B+ and anode resistor & maybe cathode resistor. Will keep you posted if I find any audible improvements on the current version.
Simon
That's 7Hz if you never connect the line stage to anything and the cap only sees the 470K grounding resistor that is part of the line stage. If you drive a power amp with a 100K input resistance (pretty high these days), you'll see a -3dB point of about 35Hz. Not good. If you load it with 20K (which the 12B4 could drive), you'd see a roll-off at about 140Hz. I think 0.056uF is just too low to advise for general use in the output of a line stage, as good sounding as it might be. I think 0.33uF is marginal too.
I'll have to do a little listening-test to check it out. Is it so very different for a preamp? At first I used a 0,47 cap; then switching to the 0,056 cap I didn't notice a more rolled-off low-end. My high-eff. speakers only reach around 60hz but I did a short test on my transmissionlines (around 25hz low-end) and it sounded okay - didn't do an a/b test though. Could you tell me how to calculate / measure the input impedance of a power amp? And how to insert it into the formula for calculating the -3db point of the preamp? I'll also give it a try with a pair of 5µF caps on the transmissionlines; if you guys are right the audible difference should be HUGE...
Klimon,
Even if your speakers don't go low enough for you to notice a huge difference in level with an undersized coupling cap, a too-small cap will introduce a phase shift well up into the mid-bass that might smear the "presence" and "air" that truly deep bass response can provide. Granted, the high frequency response of a small cap might/should be better. Perhaps you could experiment with bypassing a larger film cap with your nice 0.056uF Teflon caps.
Input R (call it “Rin”) is probably easier to calculate than it is to measure unless you are very skilled and equipped to make these test measurements. If you will post a schematic of the front end of your power amp (or whatever the 12B4 will directly drive), we can help you to calculate it.
Once we know your power amp’s Rin we can calculate the -3dB roll-off frequency as follows:
f-3db = 1/(2*pi*C*Rtotal), where Rtotal = [(Rin*470K)/(Rin+470K)] + 1000.
The last resistance value (1000 ohms) accounts for the approximate output resistance of the 12B4 stage and is so low compared to the other resistances that you can usually safely ignore it.
Even if your speakers don't go low enough for you to notice a huge difference in level with an undersized coupling cap, a too-small cap will introduce a phase shift well up into the mid-bass that might smear the "presence" and "air" that truly deep bass response can provide. Granted, the high frequency response of a small cap might/should be better. Perhaps you could experiment with bypassing a larger film cap with your nice 0.056uF Teflon caps.
Input R (call it “Rin”) is probably easier to calculate than it is to measure unless you are very skilled and equipped to make these test measurements. If you will post a schematic of the front end of your power amp (or whatever the 12B4 will directly drive), we can help you to calculate it.
Once we know your power amp’s Rin we can calculate the -3dB roll-off frequency as follows:
f-3db = 1/(2*pi*C*Rtotal), where Rtotal = [(Rin*470K)/(Rin+470K)] + 1000.
The last resistance value (1000 ohms) accounts for the approximate output resistance of the 12B4 stage and is so low compared to the other resistances that you can usually safely ignore it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.