Audio Asylum Thread Printer
Get a view of an entire thread on one page
|For Sale Ads|
Here is the schematic for the little 1 watt 12B4 amplifier that got me started on the road of parafeed SE. When I was designing it, I tried to keep costs low but not low enough to skimp on quality parts in the audio signal loop. I think any preamp that can put out 1Vrms can drive this thing to full power. I have received e-mail requests every 2 weeks or so for the schematic since I first put up the offer a year or two ago in the asylums. I redrew that schematic to make it look a little neater as well as to revise the parts used. Those that have built it and reported back are quite satisifed with the sonic results whether it's used for an amp driving regular speakers or for headphones. The 6N1P and 6H30Pi are in relatively healthy supply and 12B4's can be had at prices so low it's ridiculous. I think I had bought 6 pair of 12B4's for $35 when I was sourcing parts for the amp. Don't let that fool you into thinking the 12B4 is a bad tube.
For sound characteristic, it has a full, lively sound to it. The midrange for this amp is *large* and sophisticated (bass and treble are clean and solid as well). For the amps I've built since this one, I've tried to duplicate the midrange sound but I think I've come up a little short (even with 6h30Pi driving parafeed TJ300B to magnequest all permalloy tfa-2004). The 26 preamp driving this little amp really rolls my shorts. It kinda disappoints me that I spent 7 times the money in part costs on the 300B parafeed monoblocks and I still wasn't able to properly match the midrange immediacy and delicacy of this little amp. If only I could make it more powerful.....
I have a couple hammond chokes hanging here would the
Hammond 154E 20H, 20ma, 1666 ohms, 300VDC; be a reasonable substitute for the 1.8K ohms in the power supply? Thanks,
Nope. The current draw for that part is about 70mil with the 6N1P installed or 90mil with the 6H30Pi. In both cases the current draw exceeds the current rating of the choke. If I were to put a choke in the power supply, I'd replace the 510ohm resistor with a hammond 159M. This would provide about a 26dB reduction in the power supply ripple.
Besides knowing what the draw is on the power supply, how does the resistance of a choke come into play and does it have anything to do with the 510 ohm resistance value you had spec'd in the supply? Are there any rules for why you have replaced the resistor that is earlier in the chain (of the power supply) rather than later in the chain (the two 1.8K ohm resistors)?
159M 15H, 100ma, 256 ohms, 500VDC
How did you figure that out?
Well, I know what the current draw is because I measured it. I was also able to accurately simulate the circuit in duncan amps PSU2. This let me substitute in a choke and observe what what would happen if everything else in the circuit stayed the same.
how does the resistance of a choke come into play
The static resistance of the choke acts like a normal drop down resistor when it passes DC current. It will reduce the output voltage if you use a choke with higher DC resistance.
does it have anything to do with the 510 ohm resistance value you had spec'd in the supply?
I wanted to get the DC resistance value as close as possible to that value. The reason is so that the DC voltage at the output will not change a whole lot and the operating points of the tubes stay close to where they were before.
Are there any rules for why you have replaced the resistor that is earlier in the chain
No rules. Just a preference. You would have to buy 2 chokes if you were to replace the 1.8K resistor. You only need to buy one choke if you were to replace the 510 ohm resistor and you would still get the same reduction in ripple. Just being cost effective and effecient with real estate in the box.
Thanks, do you know how and where I can purchase the duncan amps PSU2 program to simulate your circuit and learn more? That way I can figure out how much of a difference in the DC voltage really does occur since the resistance of the 159M choke is 256 ohms at 100mA. I gather the DC voltage probably increases slightly but not by much due to the decreased resistance of the choke and the current the circuit is sitting at i.e. 70 or 90mA depending on the tube chosen.
The 1.8K resistor actually draws 35mil with the 6N1P and 45mil with the 6H30Pi. The current draws I gave earlier are for the 510 ohm resistor. Although, the 154E still isn't properly rated for replacement of the 1.8K resistor.
If only it could have more power…..That’s was my daily mantra with the Zen. At the end of the day 1.5w is 1.5w. Once TG hooked his Laurels up to my speakers it became impossible to pretend that I had enough power.
More power. Ok you’ll think I'm pulling your leg but I’m straight up serious. Take the output transformer off the 12B4 and stick a small good quality cap on the plate and hook it up to a 6L6 Cathode follower. Parafeed of course – autoformers is even better. The 6L6 is run as a tetrode and sits on a choke that can handle 80-100ma.....
My current amp is a 6H30 input a choke loaded 10Y and 6L6 cathode follower. It’s a GREAT amp. Before that I had my Zen amp modified to drive the 6L6 cathode follower and it sounds like a great big Zen with killer bass and much better highs the the standard amp. It had all the good things (and some of the bad). What was the bad? Well in my opinion the Zen is a great little amp but the midrange is not right (I'm being very picky here so no flames please.). Its dark and a bit closed in and not properly integrated with the music. Just my opinion.
The 6L6 Cathode follower is incredibly transparent. You hear the SV83 as it is. The SV83 as a triode it’s a good but not great tube. Lets stop fooling ourselves – you can do much better. Going from that to 10Y was a giant step up. So now I have 10Y with about 6 or so watts. You can have even more power with a KT88 output. The KT88 is a bit warmer and not as clear. After living with both I think the 10Y works best with the 6L6.
What operating point do you run the 6L6 as a Cathode follower? pentode or "triode" connection?
I assume then the output tranny is easier to design... less turns ratio.
i fooled with this briefly with triodes and never found and major gains... i too though the reduced turns ratio would be a huge plus.. but but found to make it listenable, i still needed much more of a turns ratio in the cathode then in the plate.. .so i was getting a great impedanc ematce and control, but suddenly i needed tons of drive to get reasonable power out so i abandon it...
i note that anthony uses pentodes... how does a pentode behave in CF mode????
also the whole reason i was looking into it was to get some serious power... and i think i was barking up the wrong tree... i never looked at it for a sweet low powered sound...
I stumbled into this just by fooling around one night. I never expected it work that well. What I have worked out is that the driver tube is king. Use an ordinary driver and you will have a ordinary amp. I'm aware of a couple of design on the web one use a 6SJ7 and another uses a transformer coupled 6SN7. I don't think either tube is really up to the job. The tubes I would consider are 71A, 45, 46, 2a3, 10Y - you get the picture.
I started with a triode in the cathode follower but found it did not work very well, not that much power and it would get distress very easily. The tetrode (or pentode) is another ball game. Its is big, clear and confident.
I guess at the end of the day this is not an easy amp for getting lots of power but a way to get the character of a small power triode without the power limitation....And heck its been lots of fun.
Its a pentode. Works best that way. The 6L6 is running at about 460V and 300 on the screen its pulling about 55ma. So only a 6L6GC is up to the job. Don't use metal 6L6 or 6L6GB etc.
The load is about 2.5K which is high but works very well. Very punchy and lots of great bass.
Since you are output power hungry,have you considered the 6CK4 as the output tube? I guess you wanted to keep it real triode with the 12B4 rather than strap a pentode.
Thanks for posting the schematic. I am one who requested it some months ago and you sent it to me within hours. I think its real nice and I have already started collecting parts and am ordering the 6CK4 to try as well.
Have you seen the new 60H@65mA choke at handmade?
What do you think its effect might be on the circuit?
The 6CK4 is a triode and has similar mu, gm, and rp to the 12B4. I can't use it with my current circuit because of the filament requirements.
The choke from handmade certainly seems like it'd be a good choice for a plate choke considering the ratings. I'd be curious to know where the choke's resonance point is at.
My visit to the handmade site pointed something out to me: the hammond 125E is not airgapped while the 125ESE is air-gapped. This kinda makes me uneasy to recommend the 125ESE for this application because the air gap reduces primary inductance. I think I'll stick with the 125E for now.
I think you would want the E, since you are not pulling DC through the core. The ESE is designed for series feed SE, not parafeed. The ESE has a bigger core and is optimized for greater DC current.
Thank you so much Tom. It's quite thoughtful of you. I remember being intrigued by this amplifier the first time it was published in Valve. What's that you say? A 6H30 driving a 300B parafeed? Sounds like a giant killer. Do you have a schematic of that one too, or is it still in its design stages? Thanks a lot. This will be a great headphone amp as well.
Yeah I have that 300B schematic somewhere in the depths of my hard drive. I think I need to redraw it again for parts adjustements.
Ignore my previous recommendation for the hammond 125ESE and stick with the 125E for this application.
I e-mailed Anthony Cirella of handmade yesterday and he said that the H300 parafeed output transformer he makes is of very similar design (ie 5K output impedance) does NOT have an airgap and can replace the 125E for those builders who want to venture outside of the 'budget.' Supposedly has a little better bass too.
Food for thought, of course one can improve on this cute amp till the cows come home but it won't be budget, nor will it have the Stoppleworth signature!!!
Seems like that's another nice alternative to the 125E. Kinda nice since 5K parafeed transformer prices seem to skyrocket after that.
My original intent was to build up about 10 of these amps and sell them to folks (hence the careful selection of parts to meet a budget). Since metalworking is not a strong point of mine and aesthetics wouldn't be great for what I'm able to do, I decided not to do it. I did find a nice chassis manufacturer but the cost of the chassis would double the parts costs! Maybe I'll just buy one chassis and redo the amp that I did build.
I'm actually going to get IAG out in Killeen, TX to build my chassis. They have those nice polished aluminum ones that look nickel plated but are not. Reasonable prices too. But, nothing like doing your own on a hammond.
Good thing your schematic is so simple to follow. I was also thinking of the Handmade H3008 plate choke, but its specs look very different from what you have specified so I guess it'll have to be the Magnequest BCP-15.
Let me begin by saying I am no guru but...
I think the H3008 may be more suitable here.I believe the Hammond was chosen by Tom as the choke because it was a budget part that could do the job at hand while being easily obtainable, budget being the operative word here. Its a power supply choke and not a plate choke so there are some compromises (as always)in its use as a plate choke. Tom can straighten me out here.
The rule of thumb for the choke I know of is preferably ~20H per 1K primary Z so if you are using a opt with a 5K primary you want ~100H. Thats not to say that 15H or 30H will not work, the extra inductance is supposed to work better.
There is another rule of thumb I have read where the opt inductance should be ~4X the choke inductance, so... if you have a 100H choke you want an opt ~400H (thats a lot!!).
Thats part of the reasoning behind the fancy materials in the opt's, higher inductance.
Now, having said all this I have seen more than one design where the choke is ~80H and the opt ~20H, which is the opposite and of course it works.
I breadboarded (new word!) a "parazen clone" 6n1p - Triode wired el84 with a PP opt as choke and 125e as opt and it worked quite well.
I guess with all the rules it still all comes down to trying it and listening to see how it works, just like what Tom did.
> > . Its a power supply choke and not a plate choke so there are some compromises (as always)in its use as a plate choke. Tom can straighten me out here. < <
One thing paul joppa has mentioned about plate chokes is the resonance point . Not that this should be the primary factor but another thing to consider when making the purchase decision. I've measured power supply chokes with resonances at 6khz and a couple at 4.5khz. One plate choke I had measured had a resonance around 3khz. IMO, it depends more on the stray capacitance of the choke rather than what the dealer brands it as (plate or power supply). The stray capacitance is all over the board for different makes of chokes and labelling chokes as "plate" or "power supply" doesn't make much difference. In all honesty I haven't measured the 157G's inductance curve and I've been using the BCP-15 for a few months now. I did use a 156G for the plate choke for quite awhile and the sound was about 90% of the quality of the BCP-15 in this circuit.
> > The rule of thumb for the choke I know of is preferably ~20H per 1K primary Z < <
I made my choke value decision based on the rp of the tube. A 30H choke would put the -1dB point at 20hz when used to load a 12B4. 60H would only give you -.5 to -.6 at 20hz. You have to realize that the plate choke inductance is in parallel with the tranny primary inductance when referenced to the tube. Most non air-gapped transformers provide craploads of inductance leaving the deciding factor on the plate choke. I think I measured around 560H on an all permalloy TFA-2004 at 20hz. I imagine the 125E is no slouch either.
Just keep in mind that even with a 9H choke in, this circuit still produced great sound. Going to higher choke values only optimized circuit performance rather that dramatically transforming the circuit sound. It still works and it still produces similar sound with "budget" parts. The parts were chosen for cost/performance ratio. When you plan for production, there's not much sense in spending 700% more for only a 5-10% increase. The customers can spend that money if they personally feel it's worth it. I fully encourage using better performing parts but I also encourage keeping it within reason and personal budget. You can build another project with the cost difference in a single part! There's a lot of fun and satisfaction in building a great performing amp for under $200. I could have easily spent $800 on this project by just using fancier iron. The question is: would it have been worth it?
Thanks Tom. I see your point. Thanks for the "resonant Point" link.
I think the last sentence of my former post sums it up...
"I guess with all the rules it still all comes down to trying it and listening to see how it works, just like what Tom did."
Actually, it seems the H3007 (and H3008) would make a decent choke in this circuit because of its ratings and cost. It'd be nice to have a frequency vs. inductance plot of these 2 chokes for comparison.
here's another new arrival on the plate choke scene - the one electron units, for $60., can do 48H at 75 ma and have the interesting series/parallel options.
I'll see if Anthony Cirella can post a frequency vs. inductance plot of these two chokes on his site or this site.
This post is made possible by the generous support of people like you and our sponsors: