Millett "Starving Student" hybrid amp - Page 280

That's one of the fabulous things about this place - not only are people willing to help out and answer simple questions like that, but they explain them so well.

Thanks, guys! That was perfect!

Tomb has the basics correct, and since the supply voltage is much bigger than the forward voltage drop of the LED the error introduced by his simplified calculation is insignificant. To be absolutely correct the voltage you would use would be the voltage drop you would like to see, in this case, (V supply - V LED) 48-1.5=46.5

As you can see, the error introduced in tomb's calculation is on the order of 3%; clearly insignificant, but if V supply was something like 5 volts, the error would be more like 30% which is significant.

Hey guys,

Will either the Electro-Harmonix 12AU7s or the JJ 12AU7s work with the_equalizer's modified version?

edit: sorry about the double post

Equalizer-Dsavitsk's 12AU7 mod should work with any 12AU7 and it only requires changing the resistor to the MOSFET - 390K ohms, I think? (Help me out, guys.)**

There are a couple of tradeoffs with all of them (including the 19J6!). Logistics's uses some very inexpensive tubes in keeping with the Starving Student tradition, but they only use a single triode (I think - please correct me if I'm wrong, equalizer!). That might make them more sensitive to matching instead of averaging out like parallel triodes. Probably a trivial point, though.

Dsavitsk had proposed using the 12AU7 and its double triodes in parallel like the 19J6. Equalizer actually built one. While 12AU7's are easily among the most common tubes (even new manufacture!), they tend to be more expensive - probably a $10 range is a more realistic average, but that's just my own SWAG - some can be in the $hundreds if you want a true audiophile 12AU7 (Mullard, Telefunken, Amperex, etc.). There are some that can be found quite cheap - like the 5963's.

** Edited per Lou Erickson's and Equalizer's posts. Sorry for getting it mixed up.

Dsavitsk mentioned a single resistor change, but the_equalizer posted it first. Here's the post.

The post says the change is to change R2/R8 to 390 Kohms, and a different wiring for the socket. (That makes sense, as it's got a different number of pins.)

I decided not to build this, as the 12AU7's are a bit more expensive, although not hideously.

I'm going to build a p2p of Logistic's 12SR7 based solution, cause they're cheap. It had a couple more changes, see the post for detals.

Both of these may also need bigger heat sinks, or may run hotter, too.

I don't know the differences between the internals of the different tubes, or how closely they may have to be matched. One reason I wanted the cheaper tube was so I wouldn't feel bad about buying extras. I also liked the metal tube, and am considering taking them to work, where they might benefit from being sturdy.

I ordered some turrets and a turret tool to try and see if I could lay out a turret board like Dsavitsk suggested. Might be fun. Not sure if I'll just push components around on the board or if I'll lay it out in Visio or something yet.

Sorry if my question was a little confusing but what I really wanted to know is if all 12AU7 tubes are swappable with eachother; if they all have the same circuit requirements.

So, if I were to build the SSMH with the modifications that the_equalizer posted, I should be able to use any 12AU7 tube?

Thanks for the awesome replies guys

Sorry for the delay in replying. I don't have access to the forum from the office.

getllamasfast: indeed they are intercheangable and, as a matter of fact, I built my 12AU7 version using new, current production JJ 12AU7 tubes. I also tested with 12AX7 (current production Russian Mullard) and 12AT7 (old GE ones) and they work great too; giving you different levels of gain.

TomB: actually the modified version I tested and built, based on Dsavitsk's hints, uses the two sections of the triode in parallel, just like the 19J6 version. I wanted to use the 12AU7 precisely because you can get some old ones dirt cheap, some new ones at around $15 USD or some 'boutique' NOS ones at $$$$. So while the current production 12A_7 tubes are not as cheap as the 19J6 was some years ago, at least their price is consistent and they're very easy to find.

Oh! After reading the posts again I see what happened... you were thinking about Logistic's 12SR7 version. That one indeed uses NOS tubes (beautiful octal ones ) that are still VERY cheap. As I remember it those are radio tubes that have a triode and two diodes in the same bulb; so there's no way to use the triodes in parallel.

cheers!

Yep - sorry about that. I'll call it the Equalizer-Dsavitsk mod. My apologies to Logistics. Thanks for correcting me!!

On a different note, you guys please be patient with me on the kits. Shipping goes slow during the week. I only have a few hours each night. I've sent out 8 so far, but it may go into Saturday before I get them all shipped.

I'm sure we're all excited to get kits, Tom, but we understand. They go out when they go out, and we'll be happy when they arrive.

It gives us things to look forward to.

I didn't realize the 12SR7 version only used one of the parts of the tube, and that could make matching more important. Is it possible to measure tubes to match them without a tube tester? Those are not as common as they once were - I think the guys at Radio Shack would look at me funny if I showed up with tubes and asked to use the tester.

(Might be fun to go down there and ask, though. Wonder if the high school kid behind the counter has ever SEEN a tube before?)

Matching tubes, particularly in this circuit, is really not that important. But, if you must, the easiest way to do it is to use the circuit itself. Drop a tube in and measure the voltage on the cathode and the plate. Then, find another one with matching voltages -- you have a match.

For the earlier discussion about using 12AU7's, or any tube with a 12.6V heater, you simply need to bias the mosfet's gate differently. The gate needs to be ~4V or 5V higher in voltage than the voltage you want on the heater. In the original, we use an even divider off of B+ to put the gate at a little less than 24V, which in turn leaves the heater at about 19V. So, for the 12AU7, you want the gate to be at about 17V give or take. So, you just need to adjust the resistor divider.

Several additional things to think about --

1. Once the heater voltage is set, the heaters will draw whatever current they want. A 12AU7 will draw a reasonable current, but put a tube with a high heater current in the circuit and it can cause problems (namely heat and too high a draw on the power supply).

2. The mosfet is dropping an additional 5V which means an additional 3/4 a watt. Not a ton, but enough that you might need to beef up your heatsinks.

3. Adjusting the Rp and Rk (the resistors connected directly to the Plate and Kathode) would be prudent. To do this, you'll need to get out the datasheet for your chosen tube and draw a load line. This is easy to do, there are tons of web pages with basic instructions, and it will be a good learning experience.

-d

The information found on this forum astounds me. I'm not quite finished with school but I never would have learned most of this in the curriculum (especially the tubes, obviously).

Yours has shipped, by the way. Here's the cutting edge these days for tube testers (and perhaps the only consistent source for a tube tester):
tube tester, great deals on Business Industrial, Electronics on eBay!

There are a couple of companies that make some modern, computerized/digital tube testers. I don't have the links handy, but they run into the $thousands of dollars, unfortunately.

This guy is an outstanding resource on tube testers and I've learned a lot from studying his website:
Tales From The Tone Lounge; The Idiot's Guide To TubeTesters!

Just an FYI, but I use the Jackson 658A, sometimes the 648S - he talks about both in the link just above. Of course, there's a bunch of info on Hickoks, B&K's, etc.

Also, Jim Cross at Vacuum Tubes, Inc., Vacuum Tubes, Electron Tube, Radio Tubes, Audio Tubes, Industrial Tubes, Guitar Tubes, Capacitors, Tube Sockets, Tube Testers is the absolute authority on the Jackson 648 (he re-wrote the manual), another very popular tester. He's quite a resource on testers and tubes as well and is a prominent member of the Tube Collectors Association. The current president of that organization, Ludwell Sibley, has the authoritative book on "Tube Lore."

1.- Indeed that was another reason for choosing the 12AU7: it's heater is spec'd to draw the same amount of current as the 19J6 heater. A friend of mine was experimenting with the 6J6 (a 6 volt heater version of the 19J6) and tried dropping the extra voltage with a resistor, but it's just so much extra power because the 6J6 heater draws 450 mA.

2.- In my build, with bigger heatsinks, the MOSFETs get so hot you cannot hold your hand on the heatsink for more than a couple of seconds.

3.- I really wanted to ask you about this Dsavitsk, since I've racked my brain (only in my free time, though) trying to do this but I've been unable to. How can you draw a load line for a 12AU7 using a 48 V B+ when the curves for the tube end with a 50 V B+ ??? I easily did that for the 12AE6 version that I 'designed' on paper (using a 30 volt HP printer supply), because the 12AE6 tube was designed to work at low B+ (I think it's rated for 27V max) and so it's curves include voltages of 50 and lower in 5 volt decrements.. but for a 12AU7 ? In fact I mentioned in my 12AE6 version post that if only I could find a way to choose proper values for Rk and Rp for the 12AU7, a version using the 30 volt HP supply could work too.

cheers!

Well, you should be able to make a reasonable approximation from the 12au7 curves. Even without detailed curves, you can make an educated guess -- for instance, pick an operating current, say 1mA. Then, try to split the voltage between the load and the tube -- thus, 24V/1mA leads to a 24K load. If you look at the curves, ~-2V grid bias is the half way point on this line, so that puts the cathode resistor at 2V/1mA = 2K.

You may need to adjust from there, but that gets you pretty close. You can also look at the 12U7 curves which are pretty similar at low voltages (12U7 is to the 12AU7 what the 6GM8 is to the 6DJ8).

Oh, remember that since the tube sides are || that you need to double the current scale on the plate curves.