|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
Don't know if the Lundahl 1660 has enough current though, cause that is one I'm thinking of.Don't know what current should be running through the C4S and Gary Pimm's CCS..anyone have a tip.
By the way found a great source for 6bx7..and other tubes. Found them throug e-pay. Bought some sockets from them allready...service is good...The tube center. $6.00 a pop..good price right?
http://www.thetubecenter.com/tubelist7.html
Follow Ups:
How bout subbing the 300b with a 813???
BAs,why the transformer after the 6BX7?
Hi 7N7,Apparently the high grid resistance of the 813 makes it hard to drive?
Paul Barker and Joe Rosen suggested that an interstage makes good sense in driving the 813. Unfortunately my knowledge is to little to make any real sense of it all. And I am basically being pushed and pulled towards one or the other direction (pun intended).
I think the design as it stands it to complex for me. I need to get back to some basic design. Whadya think?
Cheers,
Bas
Hello Bas.Grid resistance? I don't understand.
I use a 6BX7 to drive my 813s. As Fatbottle suggests, I have a 6SN7 DC coupled to the 6BX7 which, being a diff pair (p-p) sits on a pentode sink. The 6BX7 is cap-coupled (0.47 uF but should be a bit bigger) to the 813s. Each 813 has a 33k grid leak and there are 4.7k grid stoppers - whicxh are probably not necessary, but I ahve had enough of parasitic oscillations!
The main problem with driving large valves is the capacitance. 813 has a mu of about 8 so the gain will be somewhat less than this - perhaps 5 - I have not measured it. What I am getting at is that the consequent Miller capacitance is unlikely to be a serious problem especially if you use a valve such as 6BX7 that has a low anode resistance. My 6BX7 runs 15mA per side and uses 15k anode resistors.
best of luck
Yours is proven. I think I'll build something similar. Listen to it for a couple of months and then, go try and tweak...
just the picture
Bas,
I've been looking at taking the latest version of the Amity with the PCCS's on the ECC99's and subbing in the 813's and appropriate fixed biasing through the IT secondary. Or no IT and RC coupling. The 813 needs a bit less drive than the 300B and the capacitances seem about the same so it should work.My other option was a pentode driver like an SV83/EL84 with a shunt reg.
Keep it simple.
nt
When I was getting my feet wet with the Amity (my first try at triode design) I found it useful to think of everything in current flow, not voltages. Why? Well, voltages come from a stiff supply, something like the wall socket, and don't tell you anything useful about useful work being done. Voltage merely means voltage *potential*.An example? Pull the tube out of the socket. Plenty of voltage, but no amplification at all. It's current flow that does the work - more particularly, the *AC* current flow from input to output. This is where our friend the cathode resistor comes in ... it's what sets the quiescent current flow. Once you've established a nice steady DC flow, just remember the tube is merely a variable resistor that is controlled by the voltage modulation between the grid and cathode. That's all any tube does.
The next useful simplifying model is visualize a little AC source inside the tube. This lets you look at the AC behaviour of the circuit without getting lost in the DC analysis, which in the end, merely serves to energize the circuit and light up the tube. In fact, in more sophisticated designs, great trouble is taken to isolate the AC (audio) and DC (power) paths.
None of the models are exact, but they are nice "thought-experiments" that let you quickly analyze how a circuit really works. Just separate the AC and DC parts of the circuit, and think about how current flows through each portion of the circuit. Calculate the currents, and you might be surprised how accurately the physical amplifier comes to the model ... the Amity came within 5% of the calculated currents when it was first wired up. I was as surprised as anyone when the numbers appeared on the DVM.
Hi Red Pill,If that is the case then I have really been barking up the wrong tree!! Haven't I. Did not know it was that easy. Then I could KIS and just do a simple RC coupling and presto!! Could always slip in an IT later...
Mmm this appeals to me. I had an 6SN7 stage up here earlier. There were plenty of doubts then. And the 6SN7 could deliver about the same drive as an ECC99 could it not? So why all the calls for more drive?
So I'm pretty confused now...Well that is the price to pay I suppose if you don't study enough.
Weeeeelll won't bother everyone with one of my schema's this time. Guess I'll just have to breadboard something (when the time is right) and try it out fir miself.
Thanks again to everyone involved....
Howdy Bas,Sorry I meant the Aurora schematic, not the Amity. This is where I first really discovered Gary's PCCS's. Must have skipped over the previous threads here.
I'd planned to breadboard and try Aurora RC coupled, pentode driven, and maybe IT. I'm not sold on the IT myself as I shouldn't get near clipping, and they're expensive to try. My pentode idea's basically Thorsten's Legacy front end.
Cheers
RP
do you need a decoupling cap between the 20hy choke and the CCS?putting two constant current devices in series often causes lots of problems.
actually do you need the 20hy choke at all???
do you also want a decoupling cap on your fixed bias supply???
and finally the single dropping resistor will not drop anything unless current is being drawn, so the grid will be at the negative value of the B+ potential.
you need another resistor to ground to form the voltage divider, and actually a series resistor with a pot to ground will make the whole thing adjustable...
dave
The answer is I don't know on all counts...ok that is a lie :-)Because I'm not an EE and have little experience building these things..(I've only built 1 preamp and 1 poweramp ever, these were both basic), so I tend to overdo things in the psu. I suppose.
Won't the tubes always draw current?
So I guess I should move my decoupling caps to decouple the ccs'es.. would be one conclusion I could draw from your suggestions?
Cheers,
Bas
OK... im not an EE either :-) but getting the drawing right is the first step to proper assembly.you still need a DC blocking cap on the primary of the IT... otherwise the DCR will provide a short to ground... imagine 150V on the plate of the 6BX7 and a 500 ohm dcr of your IT, without a cap to block that DC to ground the trannie will try to pass 300ma....
onto the bias supply, follow the DC current loop, the output of the filter forms a series circuit feeding the resistor, IT secondary, then the grid of the tube... it is completed by the grid to ground resistance which is extrememly high, so virtually no currnet will flow... no current means no voltage drop across your resistor, so you will have nearly the full B+ (inverse because its a negative supply) on the grid.
you have two options (well many more)
you could place two resistors in series to ground and size them properly so the junction of the two act as a voltage divider sized to give you the proper bias voltage... say the supply is -500V... if you put a 100K and a 10K resistor in series to ground with the 10K grounded, the 500V will see 110K to ground. that will draw 4.5ma... 4.5ma across the 100K resistor will give you 450V and the remaining 50V will be across the 10K resistor so this will provide you with 50V to feed into the bottom of the IT, since no current will flow, that will be your bias.
you can play with a cap from across the 10K resistor to see if it helps.
its all about learning to use ohms law... that is the first step in sucessful design.
And now I understand the cap on the interstage part...Thanks
i'm sure its an oversight, but you are missing the DC blocking cap on the Interstage.dave
I'd shoot for 250 Vp/9mA for the 6SN7, and 250 Vp/15mA for the 6BX7. Use 820 ohms on the cathode of the 6SN7, 910 ohms for the 6BX7. Go for largish bypass caps, say 470uF. Even better is a parallel 6BX7 at 30mA and use 470 R on the cathode. I'd also use fixed bias on the 813.
...here's an idea . Hang the 6SN7 / 6BX7 HT from that of the 813 via more LC stages . I guess you are using > 800V . Direct-couple the 6SN7 anode to the 6BX7 grid and use tube diode cathode bias for the 6BX7 . It would loose a cap !...FB
That's more like it.I'd apply bias for the 813 with a fixed negative supply to the secondary of the IT, so no cathode resistor on that.
Select a current for the 6bx7 that the Llundahl is happy with, look it up on their website.
As for selecting cathode resistors of first two stages, you'd have to ask Garry but I think what is best is to predict right value for the current you intend to set in the ccs's then fine adjust the ccs for best sound.
A couple of q's though. And forgive me if I am ignorant. I did a search before asking if that is a consolation ;-)But how do I apply fixed negative to the secondary of the interstage?
and by using a diode as "cathode resistor" for the can I just insert any diode, I was thinking a 6s19 or 6080?
Hi Bas,
Looks good from here. On the negative voltage application of secondary for 813's g1, you show the other end grounded. Attach suitable adjustable source as with any set of pentodes for bias. Since the 813 has low gric risitance ratings I'd keep something like that value instead of traditional 100's of thousands of ohms for a bias supply and pot.
Douglas
a picture can say more than a thousand words..
Hey Bas,
I would try to eliminate the 33k grid resistor to the 813. I think the resistance of the secondary driving the 813's grid will provide an adequate return to earth, or negative if you are inclined to also eliminate the valve or resistor in the 813's cathode...it would require a modest negative supply, probably on the order of 50 mA if you went so far as to parallel the 6BX7 and adjust the 813's idle current that way. An adjustable negative of that magnitude is probably easy enough to do......If you're in too deep, take a bigger breath so you can hold it longer the next time...;-)
regards,
Douglas
...like it ,like it , but there will be no bias 'control' with a 6S19 strapped like this (see my new posting regarding triode bias) . For fixed bias , simply connect the secondary lead of the IT , that is currently connected to 0v to the appropriate negative supply , a stack of lithium cells may be ideal here and avoid another LT transformer . You may have problems with the dissipation of the CCS on top of the 6BX7 if the HT is high , also high voltage transistors may have to be used which usually have low Hfe (current gain) . Why not use a CCS for the first stage (or srpp etc) then resistor anode loading of the 6BX7 . 2nd harmonic cancellation should take place between the 813 and 6BX7 , but may be a pain in practice to implement adequately , but then again , thats helf the fun ! Remember that a CCS loaded stage will only be super-linear with an open circuit load . The 813 is a big valve , therefore probably has big capacitance Cin and miller ....FB
Hi all. Just thought I would stick some of my thoughts in here."a stack of lithium cells may be ideal here"
One thing to watch when using batteries for fixed bias- the operating point is sensitive to line voltage fluctuations. If you are using a transformer to generate the negative bias the amp will be less suseptable to line voltage changes. The bias will increase or decrease along with the line voltage changes.
"You may have problems with the dissipation of the CCS on top of the 6BX7 if the HT is high , also high voltage transistors may have to be used which usually have low Hfe (current gain)"
The PMCCS Bas is using on the second stage is only limited in voltage and power disapation by the choice of pentode used. Doesn't use high voltage transistors.
"Remember that a CCS loaded stage will only be super-linear with an open circuit load . The 813 is a big valve , therefore probably has big capacitance Cin and miller
This is true for "normal" CCS's. The PMCCS has the capability of using the bottom Mosfet as a mu folower. If Bas drives the interstage tranny from the mu output of the PMCCCS he won't have any problems with driving the output tube. With the Ra of the 6BX7 around 1.3K the output Z of the mu folower connection should be less than 100 ohms.
Gary
I agree that a CCS on the input stage would be beneficial (low noise , low distortion etc) but on the driver (resistor loaded) I would have thought that the operating point should be optimised so that some cancellation takes place between driver and output stage leaving just a little bit of second harmonic . Just a thought , or am I completely wrong ? 813 is a tetrode , therefore strapped as a triode will generate shrill 3rd harmonic albeit a little , with a little 2nd from the driver this should 'dissappear' somewhat .
I know what the purpose of the negative supply is....
cause I'm a little out of my depth here..(understatement)Regards,
Bas
...negative bias supply is for the 813 grid , simple as that . Low noise and good stability are important , after all crap / noisy unstable biasing will ruin a quality setup , hence my recommendation for a stack of batteries for the bias , but a separate variable negative supply should do the trick . You will have to check the curves for your required operating point and read the bias value from there .
Hint : when you're out of your depth technicaly , 'tread water' with the knowledge you already have . As this is quite an ambitious project I would recommend breadboarding a single channel rather than building a chassis and EXPERIMENT . Gone off 6C41C ? I've already built mine !...FB
how bout this then...
That is dual bias. No need for any auto bias element in the 813. Scrap the 6080, ground the cathode of the 813 and yes yuou have the right idea about applying a negative voltage to the grid of the 813 through the IT secondary, that's what I meant. I suppose you are goinmg to need adjustment up to -80v. If you could afford to do it in Lithium's all good and well, but grid will draw some current, check voltages often.Very ambitious amp, but good luck with it all.
Hi Paul,Thanks for the suggestions. Yes, this amp is ambitious..curiously the reason for this is actually my lack of knowledge and incorporating the suggestions from all the aa members. :-) Normally I go for ultra simple amps.
Cheers,
Bas
I would have expected a newish person, incorporating the mix from this place would end up in an assylum, but you are describing something that should work. Scary.
It was PP wasn't it? I would love to see a schematic. If you need a space to park your design somewhere I could put it up on one of my web spaces...Cheers,
Bas
...no the 6C41 amp is resistor loaded parafeed SE . Battery biased 7AF7 input stage , LED biased 7N7 driver . 0.1uF teflon couplers and before anyone asks , I don't have any spare ! In the picture notice the array of resistors behind the 6C41 which are it's anode load . HT is 350v , 6C41 anode sits at 200v , the amp gets hot , it's inefficient , the bass is excellent and the top end is crystal clear . Flea powered but cheap as chips , toroids as outputs , only a rugged , 240mA 400V supply is required . Don't even think about trying this with an 813 , the HT required would be collosal ! Low Ra all the way for this topology . .
Why not build something similar yourself if you already have the parts ? It does not sound at all bad once the parafeed cap is 'optimised' (note paralleled 1uF orange caps , I only had an 8uF polyprop) ....FB
Hi Fatbottle,What output toroidals did you use?
230V:24V 50VA units , seem to do the trick
Have you had any experience with using 115+115:50+50 toroidal
for IT purposes.
i'm considering using one to drive 4 x triode connected 807
in push pull
input to IT will be 5687 thoughts comments ??
...no experience but why not try it anyway ? I think that you can obtain such toroids as small as 50va , not smaller , but still very cheap . Worth the risk if you ask me . If you parafeed with the primaries in parallel , this should more or less be a 1:1 . Good luck !
This is just a future project.. But because my knowledge is trivial. I post here to let others review, suggest and improve etc. And let the whole thing sink into my subconsciousness ;-).
This post is made possible by the generous support of people like you and our sponsors: