• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

6146 in the Simple SE

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
OK, I was sitting in the shack Sunday afternoon, half listening to the Collins net, and the more I stared at a set of 6146A's sitting on top of my Ameritron, the more I thought about putting them in the Simple SE.

I mean, if 6DZ7's can go in the Simple SE, how hard can 6146's be?

In researching this idea, it seems there is a dearth of information on (successfully) using the 6146 in single ended audio, and most of the lack of success falls squarely on the lack of a proper screen supply. My Simple SE has a separate winding that I can use for screen voltage, so I figured it was plenty doable on my particular amp.

Indeed, after studying the pin out of the 6146 types, and the layout of the Simple SE board,
I came to the conclusion that I could drop in 6146's, and drop in a separate screen supply,
without cutting any traces on the board.

The gist of what I have done is this:

1.: I lifted the return line of the choke from the board, and lifted the hot side of R4 , freeing up the tube rectified high voltage, and isolating the remainder of the power supply bus. Having already installed banana jacks for anode caps in a previous overhaul of the Simple SE last year, I removed the OPT connections from the Simple SE board, and wired the B+ directly to the transformers and thence to the banana jacks to complete the plate connection. A 100uf capacitor completes the power supply, so I have CLC, where C is on the Simple SE board and L is on the chassis in the conventional manner, and the final C is installed at the junction of the two leads going to the primary of the OPTS.

I ran a jumper from the new high B+ to the hot side of R4 previously lifted, to restore the original relationship between the B+ and heaters.

This completes the B+ supply.

2.: I built a separate screen supply on a small piece of perf board, using a full wave (2 diode) rectifier in a CRC configuration, using a 22uF first C, a 150 ohm resistor, and then I returned this line to the LI pad I had previously lifted, thus utilizing the 47uf cap on the Simple SE board as the final C in the CRC configuration.

Unloaded, I was seeing 247 volts on the new screen supply.

This completes the screen supply.

3.: Because the 12AT7's will now be running off this lower voltage screen bus, I jumpered R14 and R24 to make sure the 12AT7 does not starve for current.

4.: I figure at some point I'm going to have to build a negative feedback loop, so I removed R10 and R20 and reinstalled them on the solder side of this board to make it easier to implement NFB in the future.

5. Because I had some concern that all of this would not work and I would toast a new set of 6146A's I decided to step back and start the amp with cheap tubes instead of the 6146's.

Once the plate connection is taken care of, the material difference between running 6BQ6's and 6146's, at least as far as pin out is concerned, is that the 6146 has its screen on pin 3, and the 6BQ6 has it's screen at pin 4. Because the Simple SE board brings both pin 3 and 4 to a separate isolated pad, it is a simple matter to wire the new screen supply to pin 3 or 4 as required.

6: I installed one 6BQ6A, put in a 560 ohm bias resistor, crossed my fingers, and powered it up.

The amp worked, surprisingly (!) but hummed intolerably. After a few minutes of colorful language, I stuck some scope probes in the empty 6BQ6 socket, and it looks like the hum is on my screen supply. I added another 100uf across he first C in my CRC screen supply and this did not materially affect the hum.

I reviewed the wiring and don't see any obvious ground loops, but can't rule that out. The input source is a battery powered ipod, so it's hard to see how that could ground loop with the Simple SE.

7.: I lifted my new screen supply line from my new screen supply board, and used an external DC supply to supply screen voltage, using my single point ground in the Simple SE as the ground reference. The hum diminshed dramatically, there is still a bit, but this, in my mind, confirms that my screen supply is the source of the hum.

8.: With 360 volts on the plate, 215 volts on the screen, a 560 ohm bias resistor, in pure pentode mode with no NFB, the 6BQ6 sounds surprisingly good.

9.: It was quite interesting to watch the screen grid action with the variable DC supply. The tube comes alive at about 200 volts on the screen, and from there to about 270 volts, there is about a 20ma swing in the plate current. Because the 12AT7 is still on this bus, it may have some contribution to this.

I'm open to suggestions as to how to tame this hum. I thought about adding some bypass caps
across the diodes on my screen board, but past that, I really don't know where to go.

It looks like if I can tame this hum, 6146 and the like, including 2E26, is plenty doable on
a Simple SE board, easy actually. I could move the jumper over to pin 3, and be running a
6146 right now.

Win W5JAG
 
Last edited:
Next comes an 813, right? Then on to 4-400A.

813's would be cool and I have a 10 volt 10 amp filament transformer for a pair of them, but a good plate transformer has eluded me thus far. MFJ has a transformer that runs 800 volts at 0.5 amp and it's only $47 but I'd rather scrounge a transformer from a hamfest for a few bucks.

I was at a hamfest this weekend and saw a really nice homebrew amp with a pair of Eimac 4-1000A's :eek: in it, but I was in my G8 and needed a truck to move the power supply - didn't feel like making a three hour round trip back to the house to get the truck and return, but I sure would have liked to have had that power supply :) .

Win W5JAG
 

Attachments

  • Simple SE 6146 wide.jpg
    Simple SE 6146 wide.jpg
    303 KB · Views: 1,655
  • Simple SE 6146 closeup.jpg
    Simple SE 6146 closeup.jpg
    226.4 KB · Views: 1,555
didn't feel like making a three hour round trip back to the house to get the truck and return, but I sure would have liked to have had that power supply

I have at least two, maybe three Motorola base station power supplies liberated from the Orange County (Florida) public safety system when they were upgraded. They put out 1500 volts at 1/2 amp, but weigh 75 pounds each. The transformers are of the ferroresonant variety so they make a buzzing noise. Probably too loud for audio, but could power a nice ham radio amp. Free to a good home, just figure out how to move them. I have been trying to give them away for a while now without success, I guess that nobody builds big HF amps anymore. They will probably got to scrap when I get down to the end of the warehouse. Large object with huge transformer on the left:

The 833A SE Amp Prototype

Next comes an 813, right? Then on to 4-400A.

Reality finally reeled me in from the extreme amp building binge I was on several years back. I still have the 813's but I sold the 4-400's and the 4-1000's. I still have several 833A's though, interesting trades?

There are a few 6146's sitting around my lab, I was planning to torture them in a P-P amp, but who knows........
 
I made substantial progress tonight, including resolving the hum problem. The amp is now as silent as when it was triode wired with conventional audio tubes.

I started by going by my warehouse after work and grabbing up a half dozen or so crusty looking 6146's and 2E26's to use as test mules.

After sleeping on it last night, I decided I could not rule out ground loops and so I redid some of the wiring in and out of the screen supply board. I was also less than satisified with the CRC circuit. I found a small 2H 100ma choke in the parts bin, so I rebuilt the screen supply as a CLC circuit, with a 10uF first C, the 2H choke, and the final C being the 47uF cap on the Simple SE board. I also paralleled each diode in the rectifier with a .1uF cap. I didn't separately test each revision, but somewhere in these changes the hum was banished. Now I have to find room under the chassis for the choke.

There were a few false starts along the way. As luck would have it, the first 6146 and 2E26 I tried, out of the bunch I brought home, was bad and went into runaway after a few minutes. The front panel ammeter is great for catching this. The bad 6146 had gone from 50 mils to 200 mils while I was putting tools away. I would have never caught this by ear; without the panel ammeter, the smoke would have come out of something.

Now that I think I have the hum issue solved, I can start looking forward and one of the things I am considering is whether I want to regulate the screen supply. It seems that having a separate winding on the transformer, its own rectifier and CLC filter, plus being actively loaded by the CCS for the 12AT7, it ought to be pretty stable. I know its stiff. It's as independent of the plate supply as it can get - I can't see plate current variations materially affecting my screen voltage.

On the other hand, I am not entirely certain I want to leave the 12AT7 on the screen supply. Basically, all amplification ceases below 200 volts on the screen line. I am not sure if that is because of inadequate voltage on the screens, or inadequate voltage to the 12AT7 stage. I suspect the latter. Leaving the two tied together is OK if the optimal screen voltage is above 200 volts, but at this point I don't know if that is true or not ( although I think it probably is ), as I haven't begun to tweak anything for optimum performance. I just wanted to stop the hum and make sure everything was stable.

Anyway, substantial progress was made tonight, and I think this will be a worthwhile learning experience. Even if I don't use the 6146's or 2E26's, 6DQ6, 6BQ6, 6889, and 6384 will all benefit and be much easier to use in practice once the screen is freed from the plate supply. There may be other sweep tubes that can be dropped in without a lot of changes. I certainly never expected to drop 6146's in, so there may be others out there to play with with minimal difficulty.

I even found a 4-400A when I was rummaging around for the 6146's.

Win W5JAG
 
You wanna go triodes? I have some 3-500Z that I don't use. Actually they are spares for my HF linear amplifier so I should keep at least a few. I even have a spare amplifier in the garage. You never know when you will need a spare amplifier, right? I mean, a pileup on 20 meters needs one!
 
It's been back in service the better part of this evening, the 6146's in it right now are probably pretty crummy; of the bunch of pulls I dragged home the other evening I wound up throwing all but the two that are in the amp now, away, as all of the others wanted to run away.

While it was interesting in a Rube Goldberg sort of way to have half of each power supply on the board and half off the board, once I determined I needed to run the 12AT7 off the full B+, I discarded the split rail idea as being without merit, and reversed those changes I made to the Simple SE board.

The bottom line is that pins 3 and 4 of the octal socket need to be completely isolated, so that the appropriate component / voltage can be connected accordingly depending on the tube in use, and because these pins are brought out to isolated pads on the Simple SE board, they are easily isolated.

For the 6146 / 2E26 family of tubes, pin 4 has to be left externally unconnected as it is internally connected to the cathode of the tube, pin 3 needs correct screen voltage, and plate voltage goes to the plate cap. That's basically it right there, so what looked like an easy project was actually a lot easier once I got a couple of evenings into it.

I wound up with an LC screen supply. CLC (10uf, 2H, 33uF) was settling in at about 245volts loaded, which is too much for 2E26 and 6BQ6, and at the upper limit for 6146 and 6DQ6. LC sets at about 215 volts for 2E26 (above the rating, but they seem to handle it OK) and 205 for 6146. I made the LC from the CLC by moving the hot lead of the 10uF cap to the other side of the choke, so it is easily reversible if I want the higher screen voltage for traditional audio tubes.

Presently the amp is running 400 volts on the anode, 205 volts on the screen, and with a 510 ohm cathode resistor, this gives about 45 mils plate current / tube. Even for old tired tubes, these are loafing. I have the plate load set at 5K, and cathode feedback enabled.

Subjectively, the amp is very dynamic, with loads of power reserve. RCA often described these tubes as compact powerhouses in advertising, and the description fits. I feel it would benefit from a global negative feedback loop as the frequnecy response sounds tilted to the high end of the spectrum to me. On the other hand, my wife thinks it sounds great. Usually, I am the only one to listen to music through the amp, and I do that after everyone has gone to bed, but tonight she spent several hours listening to it, and made several spontaneous commments about liking the sound of it, so go figure. This was after I went into the source setting and set the treble about 6-8 db below flat, however. I expected distortion to be audibly higher than before, and it is, but subtly so and not bad at all. It doesn't really sound like THD to me, more like IMD I would say as there is a subtle loss of clarity in pentode mode, for lack of a better description, as opposed to triode mode. I have 96ish dB speakers and it can flat shake the walls with power to spare now, so I'm Ok with that tradeoff, at least for now. I have quite a bit of pressing business in the forthcoming month, and since my wife seems to like the sound of it now, I'll probably leave it mostly alone for awhile.

I didn't take any pictures tonight because I exploded a cap last night and still have mess on the bench and on the amp to clean up. I'll try to take some this weekend of the temporarily finished project.

If one has to go out and buy 6146's I don't think this is really a justifiable project as there are plenty of traditional audio tubes to cook off, but having them on hand, and all the other components, my only cost is a few evenings of my time. If I decide to keep the 6146's in it, I'll probably pull these pulls, and put in some new 6146B. I need to save my 6146 and 6146A's for my transmitters, as they can't handle the B's and I have no other use for B's.

I've glanced at some pin outs of some of the other popular tubes and it looks like now that I'm this far into this thing that 6BG6 could now be accomodated with a couple of jumpers and a cut trace or maybe two. I'll probably leave that for someone else, although I do have some nice Westinghouse 6BG6GA's that would look good on the amp ....

Win W5JAG
 
Here are some top and bottom pics of the completed conversion showing the installation of the screen board. The thin orange wire is the screen wiring; the large orange wiring is for the cathode feedback.

Most are using 2.5K to 4K for the plate load on these tubes. I went in the opposite direction and settled on 10K as there is power to spare and I wanted to minimize distortion. I am using the 2.5K tap to maximize the cathode feedback.

The subtle IMDish distortion I was hearing was yet another bad 6146. A fresh set of 6146B's fixed that.

I have 2E26's in the amp now. RCA did publish data points for 2E26 in class A audio use: 250 volts plate, 160 screen, and -14.5 grid, IIRC. Right now I'm running 400 plate, 220 screen, and -17.5 grid, about 14 watts.

There is a short snap of oscillation on startup when the main DC supply has a cathode type tube rectifier as the plate lags the screen which is on a solid state rectifier. A directly heated rectifier solves that issue and makes a quiet start.

All in all, a fun conversion project.

Win W5JAG
 

Attachments

  • screen board.jpg
    screen board.jpg
    358.4 KB · Views: 1,018
  • Simple SE 2E26.jpg
    Simple SE 2E26.jpg
    247.3 KB · Views: 1,037
Not much work on this project over the last few days.

I finally got around to changing the grid resistance. 220K is probably too high, but I wasn't sure what the "correct" value is. I settled on 100K and installed a 180K resistor on the solder side of the board in parallel with the 220K on the screen side. AFAIK, the 220K grid resistance caused no adverse results, but I changed it anyway.

As these are 175 MHz RF tubes, stability kinda concerns me in this application, but it got a continuous 13 hour run on Tuesday with 6146A's in it at about 27 watts / tube with no issues so it seems stable, even when deeply heat soaked.

I'm thinking about going to a fixed bias setup, but that is not in the immediate future for this amp as I need to get some smaller parts to get under my 1.5 inch chassis. I can't find anything that even suggests using cathode bias for these tubes.

I've been using both 2E26 and 6146A and B and am developing a preference for 2E26, at least in this amp.

When the subject of 6146's comes up on the forums, a lot of the responses seem to be along the lines of just use a circuit for an 807. I'm not sure what the basis is for that, they are really different devices.

Win W5JAG
 
You don't have to have a separate screen winding on your power transformer to get a low(er) voltage screen supply in your amp.

If you have room under your chassis, you can put a couple of small filament transformers back to back like this:

:117||6.3=6.3||117:

and then rectify and filter it in the usual manner. You should be able to get 175-200 volts DC easy this way, and that will be about ideal for 2E26 and small sweeps like 6BQ6 and 6DQ6.

One RCA data plate ( the preliminary sheet, I think) states that 6146 can be run at up to 400 volts on plate and screen when triode wired. I have not tried that yet.

Some 2E26's flouresce like crazy, and look really good at work.

Win W5JAG
 
.... Any sonic comparisons to more conventional tubes like 6v6's, el 34's etc?

Can't really say, I'm not an audiophile - I did this mostly for fun and to learn a bit.

I intend to build a better screen supply and a fixed bias supply to run them closer to optimally. A good NFB network would also help; I've been pretty busy but have tinkered a bit with that, nothing worth posting up so far.

That said, even in the current setup which is crude, I see no reason why 2E26 / 6146B cannot sound as good as anything else run in pentode mode.

I didn't think much of 6BQ6* and 6DQ6* when run as triodes, having now also played with them a bit as pentodes, I have a much more favorable impression of them.

2E26's are cheap, but 6BQ6's are even cheaper.
 
I am still running 6146B's in the amp, and have a few more observations and have made a few changes since the last post.

All 2E26's and 6BQ6 types I have tried seemed stable in the amp, but at least one 6146B example and one 6DQ6 was not. Of these, they were so close to the edge of instability that just moving a tube from one socket to the other would be enough to start or stop an oscillation.

Working through this, I changed the 100 ohm parasitic supressors in series with the screens (R18 and R28) from 100 ohm to a 1.2K device (nothing special about this value -I had them in stock and saved myself a trip to the parts store).

Regarding feedback, right now I am using substantial plate to grid ( Schade? ) local negative feedback around the PA tubes implemented with a 100K resistor from the plate of the PA tube to the plate of the 12AT7 section on that side. This is in addition to the cathode feedback that is enabled. The pentode shrillness is gone, and it still has that "tube" sound, for lack of a better phrase, albeit in a more subdued manner.

It seems to be much better behaved now; the snap of oscillation on start up that required a directly heated rectifier is now gone. It's very quiet - with 96db ish speakers, in the absence of an input signal, I can't tell the amp is on, even with my ear directly against the speaker.

I'm still using cathode bias; for now I don't have the right parts for a fixed bias setup that can fit under my low profile chassis.

Win W5JAG
 
Regarding feedback, right now I am using substantial plate to grid ( Schade? ) local negative feedback around the PA tubes implemented with a 100K resistor from the plate of the PA tube to the plate of the 12AT7 section on that side.

I have tried this too with good results. But the last time I mentioned it in another thread I was heavily criticised for suggesting that Schade would work with a triode input tube. Several of the big names here explained that Schade wouldn't work with a triode and the worst possible circumstance would be a triode fed by a CCS. I decided to just shut up and listen to my amp.

If you have nothing better to do, try connecting the Schade resistor to the gate of the CCS chip. I needed a bigger resistor, 470K maybe, I don't remember. Warning don't try it if you don't have any spare chips. I blasted one in half and I don't know why. Replaced the chip and it worked.

Just plugging a 6DQ6 into the SSE resulted in an exploded cathode bypass cap because the bias voltage went up to about 65 volts.
 
It seems that a CCS loaded driver is very well suited for this type of feedback. I have enough extra CCS chips that I can explode a few and I'll try running it through the CCS when I have a few hours on the bench with it.

I was fixated on global feedback and fooled around with that with mediocre results. When I tried the Schade, I was taken aback by just how well it worked in this application.

Win W5JAG
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.