I used a common DC supply for the filaments in my TubalabSE amplifiers. DC bias current is measured by reading the voltage across a 10 ohm resistor in the B+ lead of each OPT. This obviously floats the meter at 350 volts, and requires careful attention. It is also why I don't recommend this amplifier for beginners. I wrote a set up procedure that uses 3 (or more) cheap mulitmeters connected into the amp before power up, thus avoiding probing around in a live circuit.
This tradeoff may be simpler than 2 more transformers.
The procedure from the TubelabSe manual is here, scroll down the multiple meter approach:
http://www.tubelab.com/AssemblyManualTubelabSE/MeterUse_TSE.htm
This tradeoff may be simpler than 2 more transformers.
The procedure from the TubelabSe manual is here, scroll down the multiple meter approach:
http://www.tubelab.com/AssemblyManualTubelabSE/MeterUse_TSE.htm
Fuling said:
Once upon a time Bob Danielak built an amp with a hi mu transmitting triode in grounded grid configuration. That was when you could get a variety of tubes under the Svetlana name; I'm guessing he used the SV572-160 which is (was) similar to the 811A.
Anyway, what he did was drive the cathode through a stepdown transformer with the primary in the plate circuit of the driver tube. What is kinda cool about this setup is that the two DC currents cancel each other to some degree reducing the need for a gap in the IT. In fact, if the ratio of bias currents is the same as the stepdown ratio then they nearly cancel completely.
It seems like a solution that could actually work pretty well. The change in load is much less drastic as the output tube grid starts conducting. The bonus is that the power from the driver tube is added to the output tube power and appears at your speaker terminals.
Don't know what's happened to "rocket boy" Bob. Haven't heard from him in a long time.
-- Dave
qq said:
If you want to play with A2 without high voltages and big power supplies, there's the little 19. It's a directly heated dual triode that clearly needs to be run with positive grids. It has a somewhat higher mu than most of those little DHT's, which is what makes it interesting to me (could use it as a driver in a 2 stage amp with all DHT's... except that it takes some power to drive it.
)Both filaments are connected together, which limits it's use a little. Could just parallel them and call it one tube...
They show up on eBay from time to time.
-- Dave
Some "zero bias triodes" like the 3-500Z, the 833A, the 838 and even the 811A can be operated entirely in the positive grid voltage region if the plate voltage is low enough. The tube will draw some grid current all of the time, but the grid current draw is not constant, so the driver must exhibit a low output impedance to avoid distortion. The sharp impedance transition around zero grid volts is obviously the worst case. A well designed mosfet follower can deal with this since the output impedance when sourcing current is the ON resistance of the fet (a few ohms) plus the dynamic impedance of its drain supply ( a few ohms). For those who wish to avoid sand contamination, use a triode wired sweep tube for the follower.
I did some experiments a while ago (5 or 6 years) with the 811A and operation with positive grid bias. I never got the sound that I was looking for, and went on to other designs. I have planned to try this again some day with some other tubes and topologies, but haven't found the time yet.
I have also experimented briefly with screen driven sweep tubes (same basic drive circuitry and positive bias requirements) and found (again) the best measured performance and sound quality comes from a mosfet driver. I have a P-P amp operating on the edge of class B (5 mA bias current) that puts out 80 watts from a pair of 6AV5's. Distortion is under 3% at 50 watts WITHOUT any feedback. Pictures are at the bottom of this page, no schematic yet:
http://www.tubelab.com/6AV5.htm
I did some experiments a while ago (5 or 6 years) with the 811A and operation with positive grid bias. I never got the sound that I was looking for, and went on to other designs. I have planned to try this again some day with some other tubes and topologies, but haven't found the time yet.
I have also experimented briefly with screen driven sweep tubes (same basic drive circuitry and positive bias requirements) and found (again) the best measured performance and sound quality comes from a mosfet driver. I have a P-P amp operating on the edge of class B (5 mA bias current) that puts out 80 watts from a pair of 6AV5's. Distortion is under 3% at 50 watts WITHOUT any feedback. Pictures are at the bottom of this page, no schematic yet:
http://www.tubelab.com/6AV5.htm
hacknet: You will need two isolated supplies if the tubes are run in SE. You are also going to need some large heatsinks and insulating hardware for the lm338s. One other idea I had was using two switching buck regulators to get a clean supply and they'll barely get warm. The only problem is they need careful PCB layout. Ripple is a non issue when the switching frequency is 400Khz 
I went with switching supplies after pricing filament chokes, both were less than the cost of one choke.
I went with switching supplies after pricing filament chokes, both were less than the cost of one choke.i was looking through the national catalog and it pointed me to th lm3489, 93% eff which sounds good and quite low component count except for the inductor which is supposed to be a 22uH. i dont know where to find such an inductor.
is noise critical in such an operation? would the amp work with a little more than usual noise on the filament? if it was say at 500khz?
is noise critical in such an operation? would the amp work with a little more than usual noise on the filament? if it was say at 500khz?
Try using the online simulations on national's site. You input the specs you need and it chooses the part numbers and values. They even have a service that will build the prototype boards for you. Not sure about their prices though...
Some tube people dislike switching supplies but they do have their place in certain applications.
Some tube people dislike switching supplies but they do have their place in certain applications.
I have everything (except for a few pictures) ready. I must be out of town for the next two weekends (maybe 3). The web pages will be updated in the next two weeks with all the info as I have time. I will probably have to use Ebay for initial board sales, since I haven't figured out how to set up a web shopping cart yet.
I may understand vacuum tubes, RF and digital design, and microprocessors, but after 5 years of the Tubelab web site, I still am clueless about web site design. Up until now I have just pasted pictures into Microsoft Front Page and added some text. Microsoft no longer supports Front Page, and I have just ordered their newest "Expression Web" I have to build the web site over from scratch. This will take place over the the next several months. The current web site will remain in place until the new one is ready.
I have been using the Priority Mail envelopes from the US post office for low cost shipping to the US and some other countries. There are several countries where this service is not available. I have not found a suitable shipping solution for those countries. $32 for Fedex shipping is not acceptable for a $40 board.
Now back to the current thread. It is possible to use a switching supply to light up DHT's. I used one of National Semiconductor's "Simple Switchers" to power some 2A3's from a schottky rectified 6.3 volt filament winding. Put the rectifiers, filter cap, the switcher, and its output filter all on one small PC board. The switcher's output filter must be made with low ESR capacitors. I know that the tube world frowns on ceramic capacitors, but they are the hot ticket here. There should be no ground connection anywhere in the power supply, it should all float. The positive and negative output leads should BOTH go through a common mode choke before reaching the 811A's filament. The only ground connection should be made at the 811A's pin. I use a pair of resistors with the ground connection (or cathode resistor) tied to the junction of the two resistors.
Yep, that's the way I use them as well. Except that I got 9V SMPS's right away for almost nothing and drop some volts in a resistor before the common mode choke. Ok, get some 10W or so more heat, but we're not talking efficiency anyway. As a side effect, the SMPS limits filament inrush current, even more of course with a series resistor.
Marcus
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.