|
Homebrew Equipment A place to show, design and discuss the weird and wonderful electronic creations from the hands of individual members. |
|
Thread Tools |
12th Apr 2011, 11:51 am | #21 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
I came upon this at Audioasylum.com.
Another way of driving a hungry tube with a catode-follower. Probably not a bad idea after all. The 3C24 will draw double the G1-current of a 2C34 and if this circuit will supply it, then it's worth looking into. Still at bit weird with the 3C24 G1 w/o grid resistor and negative supply...? rgds, /tri-comp Last edited by tri-comp; 12th Apr 2011 at 11:57 am. |
12th Apr 2011, 2:18 pm | #22 |
Octode
Join Date: Apr 2009
Location: Aylesbury, Buckinghamshire, UK.
Posts: 1,522
|
Re: 2C34 Project taking off...
I like the method of offsetting the standing current. I assume this is a mains toroidal transformer. I wonder what the quality is like?
Yes it is strange not to have a pull down, a resistor at least. BTW I've never understood the obsession with using a ccs as a pull down for a c/f. Ideally we want to keep the grid to cathode voltage constant for a faithful reproduction. With a triode c/f, the bias point will move as the voltage a-k varies. For instance (all at 2mA Ia) 12au7 Va-k 160V, Vg -8V 12au7 Va-k 260V, Vg -14V 12at7 Va-k 160V, Vg -2V 12at7 Va-k 260V, Vg -5V So a 12at7 would be the better choice here. As soon as the output grid starts to draw current, the bias point (of the c/f) totally goes out the window. Only if the standing current were many times the required grid current, and the c/f was bootstrapped will this be any way linear. You then risk the driver circuit being more power hungry and complex than the output stage. Silicon may well be the answer, or you could use a transformer. Rob.
__________________
We have done so much, for so long, with so little, we are now qualified to do anything with nothing. |
13th Apr 2011, 8:50 pm | #23 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
More interesting Class-B driver stuff here:
http://www.qsl.net/wa5bxo/driver1.html ...and this is really overdoing it imho : http://www.classeradio.com/tube_driver.htm ...oh, well they are everywhere once you start looking: http://www.geofex.com/Article_Folder...osfetfolly.htm It'll soon be time to heat the soldering iron to try out a few ideas. rgds, /tri-comp |
13th Apr 2011, 11:09 pm | #24 |
Octode
Join Date: Apr 2009
Location: Aylesbury, Buckinghamshire, UK.
Posts: 1,522
|
Re: 2C34 Project taking off...
The solutions all seem either crude or complex, with nothing in between.
Maybe the answer is to run them at a much higher ht voltage and anode load impedance, to save having to drive the grids so hard. But then the nearer you get to cut-off, the less linear it becomes. Rob.
__________________
We have done so much, for so long, with so little, we are now qualified to do anything with nothing. |
15th Apr 2011, 3:29 pm | #25 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
OK, so Philips stole my idea
Looks like they're running the negative G1 voltage (PSU output -3) around -240V if I get my Dutch right. All in all rather simple. If it's good enough taking QB3.5/750 to 1KW it'll be good enough taking 2C34 to 15Watt rgds, /tri-comp |
15th Apr 2011, 6:34 pm | #26 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
Trust the inventive Chinese to go all the way in this:
>> Look Ma, no Cap's << amplifier (3rd. page) I promise to not publish any more on the subject until I start putting the electronics together and can share my findings. rgds, /tri-comp |
15th Apr 2011, 6:38 pm | #27 |
Octode
Join Date: Apr 2009
Location: Aylesbury, Buckinghamshire, UK.
Posts: 1,522
|
Re: 2C34 Project taking off...
Seems to be a blank document friend.
Rob.
__________________
We have done so much, for so long, with so little, we are now qualified to do anything with nothing. |
15th Apr 2011, 7:22 pm | #28 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
Opens fine in my Acrobat.
It's in Chinese so you may have to add-in something to the Reader but it should suggest you to do so when you open the doc. Still, the pictures should show in any case. Someone else with problems reading this pdf? /tri-comp EDIT: I copied the 'Look Ma...' schematics from the Chinese pdf and put it through Word printing it out with Adobe PDF I also added a CF-driven EL34 PP amp, also Chinese Looks more and more like this variant will be the more successful Last edited by tri-comp; 15th Apr 2011 at 7:44 pm. Reason: Made a new PDF-file |
16th Apr 2011, 8:06 pm | #29 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
Put some values on the diagram and decided on the cascoded CCS.
There's an error in the Morgan Jones schematic (d). BC549 Collector voltage should show -134V, not -124V Any comments are welcome. rgds, /tri-comp |
17th Apr 2011, 3:28 pm | #30 |
Octode
Join Date: Apr 2009
Location: Aylesbury, Buckinghamshire, UK.
Posts: 1,522
|
Re: 2C34 Project taking off...
That last circuit looks about the best, not overly complex.
Just had another play with mine. Upped the ht to just under 400V, standing current 25mA combined, with -20V bias. A-A impedance increased to around 15K. Grids fed by 1:0.75+0.75 driver tx, driven by a 2 stage 6sn7 double triode amplifier. Will happily drive the grids to +15V. 68K added between the two grids to help stabilise the load on the driver stage. Circuit very similar to an RSGB design for a 30W modulator. Good results, clean and punchy at 2-3W, a little harsh on vocals at full power, guitar ok but struggles with bass. Rob.
__________________
We have done so much, for so long, with so little, we are now qualified to do anything with nothing. Last edited by glowinganode; 17th Apr 2011 at 3:36 pm. |
17th Apr 2011, 4:57 pm | #31 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
Rob, why do you think it gives out a meager 2-3Watt undistorted, only?
With the grids at +15Vpeak and anodes at 400V it should really whack a lot of power after the speakers. It is the curse of Class-B sticking it's ugly head out? I'm wiring the filaments at the moment; high-voltage is next. I decided to get hold of af very small 24-0-24/66mA transformer and double it to get the -110V negative supply. It's comming from RS-Components on Tuesday and I'll have all of Easter to really get into doing something usefull with this amp. During Easter the other chassis I was describing in Rob's RK34 thread here: https://www.vintage-radio.net/forum/...ad.php?t=61604 will be cut and folded. It's a more simple and traditional amp. with EL34 and Lundahl iron. It has, however a LOT more details to the chassis and will look quite differently. I'll start another thread when it becomes time to start building it. It's really not that special an amplifier; it's the chassis that should attract attention. Again, why the 2-3Watt, only? rgds, Torben /tri-comp Last edited by tri-comp; 17th Apr 2011 at 5:05 pm. Reason: Info on upcomming EL34 amp |
17th Apr 2011, 8:34 pm | #32 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
Wiring up electronics.
A few pictures to let you see the point-to-point wiring as it comes along. I thought about designing a PCB and having it made but there is really too few components to justify the costs and time. Also, if you followed this thread you'll know nothing is decided for sure about this amplifier It's a lot easier to modify P-to-P. rgds, /tri-comp |
17th Apr 2011, 10:40 pm | #33 |
Octode
Join Date: Apr 2009
Location: Aylesbury, Buckinghamshire, UK.
Posts: 1,522
|
Re: 2C34 Project taking off...
Hi Torben, that's a lovely job.
The 2-3 Watts I mention is the mean music power as measured (and hence calculated) with an rms Voltmeter. At this output it can handle peaks of two or three times this value. At higher outputs, it still sounds ok but more PA and less HiFi. This is to be expected as we are then getting well into Class B. Out of interest I replaced the RK34 with a 6080 for comparison.The 6080 gives roughly the same audible power (Class A p-p) but sounds a lot nicer. I'm trying to maintain enthusiasm for this project, but it just doesn't quite give the power output I was after. I guess I'm starting to learn why designers stopped using triodes and used (beam) tetrodes and pentodes instead. Rob.
__________________
We have done so much, for so long, with so little, we are now qualified to do anything with nothing. |
21st Apr 2011, 1:15 pm | #34 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
I calculated and drew-up the Power-Supply section.
So far the B++ for 6AN8A isn't taken into consideration; that's for later. Design is partially done with PSU Designer II / Duncan Amplification. When it comes to low-voltage design, as the 6.3V= circuit, I'm not so sure about it's ability to generate reliable data for the Voltage Doubler. Anyone with experience? Anyone with comments? rgds, /tri-comp |
22nd Apr 2011, 8:34 pm | #35 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
More electronics added today.
High-Voltage Supply is done and the Negative Bias almost done. Both Cathode-drivers mostly wired. Tomorrow I'll get a 2mm alu-strip to mount the 4 x MJE340 CCS-transistors. It should fit right across the 8 terminal lug strips above the CF's. Then it will be time to finish the CF-circuits and the few missing parts of the negative supply. Then I can't really think of any excuse to not power-up Since this is my first amp. ever, I'm learning as I'm getting along. How to do wiring and how not. Even with all good intentions and a bit of knowledge it's NOT easy avoiding AC and DC/Signal-wires crossing or even being near each other. I suppose the Hum-level will tell if I'm successful or not rgds, /tri-comp Last edited by tri-comp; 22nd Apr 2011 at 8:40 pm. |
25th Apr 2011, 12:31 pm | #36 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
The Alu-strip is mounted and so are the CCS' for one channel.
Point-to-point wiring became very dense and I don't want to re-build 4xCCS if I need to change anything, so just the two are finished. And then I wanted to share a couple of garden-pictures. Rhododendrons can be seen at the rear if you can tear your eyes from the beauty in the front rgds and the very best wishes for a Happy Easter, /tri-comp |
26th Apr 2011, 11:43 pm | #37 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
It doesn't work and it's no big mystery!
I powered-on the CF/CCS stage without any 2C34 mounted, just to be sure. It was a good idea! Actually everything worked just as it should; it's just I have no experience dimensioning stages so something had to give. If you look at the schematics for the Output/CF/CCS-stage there is something that needs to be taken into consideration. The CF-output quiescent voltage of -15V is correct; it's just not what the circuit-values will produce. 1) The B+ turned out to be more like 380V quiescent 2) The CCS is around 3mA as it has been calculated to be. 3) Look at the attached Plate Characteristics file The 5687WA is running Plate=380V and needs a VERY big negative G1 to divide this voltage between the CCS/3mA/-110V and the B++ 380V It will put it very close to cut-off and very non-linear. Something impossible to get to work. So what to do? As I see it: 1) Take the CCS to maybe 15mA 2) Reduce the B++ for 5697WA to 200V That's going to produce a lot of heat, probably around 9Watt pr. channel (Voltage-drop ~ 180V at 15+15+10+10mA = 9Watt; not impossible) Otherwise tap the powertransformer un-symmetrically with an extra Diode/Cap for a lower B++ just for the CF. 3) Each MJE340' will be burning 15mA x 95V = 1,5Watt or 3Watt pr. Channel. Also OK I hope. The transformer for the negative supply is rated for 66mA total. The TL783 will feel fine with less than 1Watt lost in it. NOW would be a really good time for your input. or NOT best, /tri-comp |
27th Apr 2011, 8:31 am | #38 |
Octode
Join Date: Apr 2009
Location: Aylesbury, Buckinghamshire, UK.
Posts: 1,522
|
Re: 2C34 Project taking off...
You could try dropping the B+ to the c/f, as you say it won't need to be so cut off.
Leave the ccs at 3mA, that's ok. Can you try ECC81 / 12AT7 as c/f instead? You say it doesn't work, can you clarify please? Rob.
__________________
We have done so much, for so long, with so little, we are now qualified to do anything with nothing. |
27th Apr 2011, 4:52 pm | #39 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
Hi Rob,
It doesn't work because the midpoint of the CF/CCS (=2C34 g1) is somewhere between 0 to +16V without a 2C34 in the socket. This node should be set on -15V to bias the 2C34 correct and it isn't. Of course I can vary the 5687WA g1 and thereby the Cathode current/voltage towards the CCS, only I can't hit -15V; it just stays too high even with g1 at -25V which is very close to full cut-off. Then I started thinking about the CF and the CCS being just passive resistors in a quiescent situation with no signal-input. The CCS is passing 3mA and needs to drop 107,8V - 15V = close to 31Kohm. The CF is passing the same 3mA and now dropping 380V + 15V = 132Kohm. That is, if the circuit was working correctly. Now, to make the CF run at 3mA with close to 400V between Cathode and Anode is reckless and not likely to work. The g1-bias should be around -30V (-45V towards Ground as the Cathode is already at -15V) Look at the Plate Characteristics and see where that would put you in a space with very cramped curves. To raise the CCS/CF-current along with dropping the B++ will put you where the curves look nicer even if it means producing heat. With B++ at 220V and CCS at 10mA the curves start looking nice but then again, am I going to be left with only around 20Vpp to swing the 2C34'? What am I missing? (....besides insight rgds, /tri-comp |
27th Apr 2011, 10:36 pm | #40 |
Heptode
Join Date: Oct 2008
Location: Aalborg, Denmark
Posts: 903
|
Re: 2C34 Project taking off...
I did more tweaking and came up with -15V at the CCS/CF common node.
See the attached pdf that shows what component-values achieved this. The CCS is increased to almost 10mA. The DC-potential for the Vg1-circuit is changed towards minus. Obviously the 8mA drain is too much. Resistors should be scaled-up by 10 but not tonight because I feel very much like As you see it takes -37V on g1 to to get the Cathode on -15V That's pretty close to the Plate Characteristics (Vg1 minus Vk = 22V) for 10ma and 380V. Now, what about an idea of reducing the B++ to 220 - 240V and setting -Vg1 to -29V? Grid-bias will then be around 14V as required for 10mA. The curves seems to be evenly spaced here for low distortion. ...or is 5687WA just not the tube for the job? rgds, /tri-comp |