This is a project I have been working on for a while and thought others might think it interesting. This is a small board for testing various circuits, all of the grounded cathode variety. The idea is that by changing a few jumpers, there are a few hundred different circuits that can be breadboarded. I am planning to use this to test drivers for a choke phase splitter push-pull amp, but it would be fine as a single ended driver, preamp, part of a phono stage, headphone amp, etc.
I used the D3a as it is reputed to be about the lowest distortion small signal tube there is. It has high mu and high Gm making it a great driver. Also, being a pentode, it will work as both a pentode and a triode allowing for trying various circuits. But, the board should also accommodate other tubes like the 6688, 7788, 7721, 7722, etc.
Below are the board design, and the circuit schematics that can be produced.
I used the D3a as it is reputed to be about the lowest distortion small signal tube there is. It has high mu and high Gm making it a great driver. Also, being a pentode, it will work as both a pentode and a triode allowing for trying various circuits. But, the board should also accommodate other tubes like the 6688, 7788, 7721, 7722, etc.
Below are the board design, and the circuit schematics that can be produced.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
If you have some of these boards made, I would like to get 2 or 3. I have been told by the "gurus" that the D3A is a magic tube. Of course I didn't have any, and they aren't too common here in the US. I bought a pair of D3A's from Germany for too much money, but haven't had the time to play with them yet. Your board takes care of the hard stuff.
I have several other tubes with the same pinout and would like to compare them all. The common 6EJ7 and 6EH7 has the same pinout except that pins 8 and 9 are swapped, but your board will handle this with a flying wire jumper. Can you post (or email me) the Eagle schematic for your board? I am having a hard time visualizing it all from the PCB artwork.
The little 7 and 9 pin test boards you did do come in handy for "testing" tubes. I haven't had too much time to use them lately, but I have "seen the light" come out of some 6AQ5's. They make it easy to wire a 7 or 9 pin tube into my octal based Simple SE board.
I have several other tubes with the same pinout and would like to compare them all. The common 6EJ7 and 6EH7 has the same pinout except that pins 8 and 9 are swapped, but your board will handle this with a flying wire jumper. Can you post (or email me) the Eagle schematic for your board? I am having a hard time visualizing it all from the PCB artwork.
The little 7 and 9 pin test boards you did do come in handy for "testing" tubes. I haven't had too much time to use them lately, but I have "seen the light" come out of some 6AQ5's. They make it easy to wire a 7 or 9 pin tube into my octal based Simple SE board.
Schematic is at http://www.ecp.cc/images/D3a_experiment_board_sch.gif
The things that look like LM317's are 10M45's. "out" is the anode, "in" is the cathode, and "adj" is the gate.
I'll probably get a few made w/o soldermask and screen to keep costs down. That way, if we split them it will cost about $35 for a pair -- this drops a bunch if anyone else wants a few. It also drops having them made overseas, but it takes a bit longer.
I haven't been able to find jumpers yet. I assume they exist as they come on the back of harddrives, but Mouser has not been forthcoming so far. Spacing on jumpers is 0.1"
The things that look like LM317's are 10M45's. "out" is the anode, "in" is the cathode, and "adj" is the gate.
I'll probably get a few made w/o soldermask and screen to keep costs down. That way, if we split them it will cost about $35 for a pair -- this drops a bunch if anyone else wants a few. It also drops having them made overseas, but it takes a bit longer.
I haven't been able to find jumpers yet. I assume they exist as they come on the back of harddrives, but Mouser has not been forthcoming so far. Spacing on jumpers is 0.1"
Hi dsavitsk,
granted, the D3a had been considered a "secret tip" over here in Europe some years ago, but since then rumours and hysteria settled quite soon. Actually, due to its extreme gm it is quite hard to tame, and also for the same reason, its factory tolerances are rather wide. If you don´t believe, have a look at the TFK D3a spec sheet .
Due to its extreme gm, to get stable operation, one is well adviced to use the somehwat self-stabilizing cathode resistor auto bias method, using *exactly* the suggested Rk values and operation points as given in the TFK sheet. Sorry to say so, but there really is not much room for any experimentation with this tube.
I don´t know where this "reputation" may have originated from, but neither spec sheet plots nor my own tracings of NOS/NIB TFK and Siemens samples suggest that. Neither in native pentode mode, nor when being triode-strapped. Granted, when considering its extreme gm its linearity is rather good, but otherwise almost any medium-to-low gm (or mu) tube can beat D3a in linearity any day, in any operation mode.
In my humble opinion, in audio the D3a just should be left to tasks that come in "naturally" by its data: Constant current sources, cathode followers and "upper" position in mu-followers (see Morgan Jones book), or even better, shunt-loaded mu-followers like Alan Kimmel stages (aka "mu-stage"). The later application is quite popular and proven for D3a over here in Europe, for those, who know how to handle and tame it for stable operation and suppression of parasitic oscillations.
Be aware that w/ D3a, parasistics (especially in CF duty) can occur at several hundred MHz - chances are you wouldn´t even detect them with a 100MHz scope. A friend of mine was disappointed by a somewhat "harsh" sound of a D3a circuit, but, according to his "scoping" of the circuit, was sure about not having parasitics going on. But there _were_ parasitics; rather easily detected by my old and trusty HP 410C VTVM, which is spec´d up to about 1GHz, IIRC. Eventually, an additional cathode (!) stopper cured the problem.
All in all, I wish you "good luck" with your D3a experimentation board - you probably will all luck, since using a PCB definitely is not the easiest way / method of choice when experimenting with extreme gm vacuum tubes, except you have designed it right from the start to prevent parasitics. Usually, it is quite more feasible to do a well-thought-out RF circuit style air- or hybrid- wiring.
Regards,
Tom Schlangen
granted, the D3a had been considered a "secret tip" over here in Europe some years ago, but since then rumours and hysteria settled quite soon. Actually, due to its extreme gm it is quite hard to tame, and also for the same reason, its factory tolerances are rather wide. If you don´t believe, have a look at the TFK D3a spec sheet .
Due to its extreme gm, to get stable operation, one is well adviced to use the somehwat self-stabilizing cathode resistor auto bias method, using *exactly* the suggested Rk values and operation points as given in the TFK sheet. Sorry to say so, but there really is not much room for any experimentation with this tube.
I don´t know where this "reputation" may have originated from, but neither spec sheet plots nor my own tracings of NOS/NIB TFK and Siemens samples suggest that. Neither in native pentode mode, nor when being triode-strapped. Granted, when considering its extreme gm its linearity is rather good, but otherwise almost any medium-to-low gm (or mu) tube can beat D3a in linearity any day, in any operation mode.
In my humble opinion, in audio the D3a just should be left to tasks that come in "naturally" by its data: Constant current sources, cathode followers and "upper" position in mu-followers (see Morgan Jones book), or even better, shunt-loaded mu-followers like Alan Kimmel stages (aka "mu-stage"). The later application is quite popular and proven for D3a over here in Europe, for those, who know how to handle and tame it for stable operation and suppression of parasitic oscillations.
Be aware that w/ D3a, parasistics (especially in CF duty) can occur at several hundred MHz - chances are you wouldn´t even detect them with a 100MHz scope. A friend of mine was disappointed by a somewhat "harsh" sound of a D3a circuit, but, according to his "scoping" of the circuit, was sure about not having parasitics going on. But there _were_ parasitics; rather easily detected by my old and trusty HP 410C VTVM, which is spec´d up to about 1GHz, IIRC. Eventually, an additional cathode (!) stopper cured the problem.
All in all, I wish you "good luck" with your D3a experimentation board - you probably will all luck, since using a PCB definitely is not the easiest way / method of choice when experimenting with extreme gm vacuum tubes, except you have designed it right from the start to prevent parasitics. Usually, it is quite more feasible to do a well-thought-out RF circuit style air- or hybrid- wiring.
Regards,
Tom Schlangen
Tom, in my measurements of the D3a, the linearity was outstanding, easily surpassing any other tube I've had on the bench. At the same signal output level that Morgan used for his 417 tests, and similar setup (CCS plate load, LED cathode bias), in triode mode, the second harmonic was at -70dB, third at -90dB, and nothing higher order was measurable. I had to double-check my measurement setup since I thought something must broken! At a higher signal level (25VRMS), I measured -58dB second, and everything else was below my measurement threshold (-100dB).
Now, I agree that it's not an easy tube to use (tendency to oscillate, lots of unit-to-unit variation, tendency to draw grid current), but the effort is certainly worth it. In my phono stage, it's dead quiet and does not significantly contribute to the overall distortion. What else could one ask of a tube?
On the other hand, please keep telling everyone that it's nothing special. That will keep the prices down.
Now, I agree that it's not an easy tube to use (tendency to oscillate, lots of unit-to-unit variation, tendency to draw grid current), but the effort is certainly worth it. In my phono stage, it's dead quiet and does not significantly contribute to the overall distortion. What else could one ask of a tube?
On the other hand, please keep telling everyone that it's nothing special. That will keep the prices down.
Hi Stuart,
first of all, I don´t doubt your measurements, since you certainly know how to do such stuff.
On the other hand, you agree on the comparatively large unit-to-unit variations, so I suppose you made a lucky pick with your samples at hand and the operation point you use, too. I (and many others) just didn´t.
I mean, considering that for the spec sheet given (pentode) op point an Ia range of +/- 14% ~ 30% explicitely is allowed for brand new out of the box tubes, should ring some bells regarding also to be expected linearity variation, when just plugging in another D3a from a different batch (in the same, unaltered circuit, of course). The OP wil move considerably, as also will the distortion spectrum distribution and hence absolute THD numbers.
Further on, looking at the "end of life" data in the mentioned TFK sheet, things get even more extreme. So, real world circuits (not talking about measurement setups) using D3a as they come out of the box and and age during use must be designed to be able to deal with such extrema in a most forgiving manner, or else one will get a "D3a rolling required - pick 1 out of 20 lottery" circuit, so to say.
Essentially, all tubes giving comparatively data and performance to D3a are beasts to deal with. And in audio, such extrema rarely are needed. Backing off somewhat from such extrema saves a lot of trouble - and money. I mean, who cares, say, if a CF gives an A of 0,97 using a bronco D3a, or A=0,95 using a tame workhorse EF184/6EJ7, or A=0,94 using a completely unsexy but made by the millions EF80/6BX6 (numbers just made up, you know what I mean).
In Europe, over the last two years or something, street prices for D3a dropped considerably from their peak. Just as some other fashionably "high gm´ers" like, say E280F, did. Just a sign that many gave up designing "real world" circuits with them, I suppose. I don´t expect falling prices for TFK EC8020, though
Uhmmm, but, well, D3a and the like _are_ special, at least for their extreme gm. But that is no hard criterium for good audio duty and sound, and often not worth the hassle. I mean, in audio we still are real men when _not_ showing off that we can tame D3a and the like, but instead design real world usable circuits that sound good using generic/average replacement tubes and keep on sounding good even after thousands of hours of parts wearing and aging.
Just my opinion, but I admit to be biased towards using unsexy and cheap and plentyful (think "TV" <gasp>) tubes and trying to get the most out of such anyway.
Regards,
Tom
first of all, I don´t doubt your measurements, since you certainly know how to do such stuff.
On the other hand, you agree on the comparatively large unit-to-unit variations, so I suppose you made a lucky pick with your samples at hand and the operation point you use, too. I (and many others) just didn´t.
I mean, considering that for the spec sheet given (pentode) op point an Ia range of +/- 14% ~ 30% explicitely is allowed for brand new out of the box tubes, should ring some bells regarding also to be expected linearity variation, when just plugging in another D3a from a different batch (in the same, unaltered circuit, of course). The OP wil move considerably, as also will the distortion spectrum distribution and hence absolute THD numbers.
Further on, looking at the "end of life" data in the mentioned TFK sheet, things get even more extreme. So, real world circuits (not talking about measurement setups) using D3a as they come out of the box and and age during use must be designed to be able to deal with such extrema in a most forgiving manner, or else one will get a "D3a rolling required - pick 1 out of 20 lottery" circuit, so to say.
Essentially, all tubes giving comparatively data and performance to D3a are beasts to deal with. And in audio, such extrema rarely are needed. Backing off somewhat from such extrema saves a lot of trouble - and money. I mean, who cares, say, if a CF gives an A of 0,97 using a bronco D3a, or A=0,95 using a tame workhorse EF184/6EJ7, or A=0,94 using a completely unsexy but made by the millions EF80/6BX6 (numbers just made up, you know what I mean).
In Europe, over the last two years or something, street prices for D3a dropped considerably from their peak. Just as some other fashionably "high gm´ers" like, say E280F, did. Just a sign that many gave up designing "real world" circuits with them, I suppose. I don´t expect falling prices for TFK EC8020, though
Uhmmm, but, well, D3a and the like _are_ special, at least for their extreme gm. But that is no hard criterium for good audio duty and sound, and often not worth the hassle. I mean, in audio we still are real men when _not_ showing off that we can tame D3a and the like, but instead design real world usable circuits that sound good using generic/average replacement tubes and keep on sounding good even after thousands of hours of parts wearing and aging.
Just my opinion, but I admit to be biased towards using unsexy and cheap and plentyful (think "TV" <gasp>) tubes and trying to get the most out of such anyway.
Regards,
Tom
Tom, for the record, six samples, all Siemens, with two out of Morgan's stash (he had baked them to try to reduce grid current), four from Frag Jan Zuerst. 1.7V red LED for cathode bias, 10mA current. No more than 1 dB variation in 2nd, and 3rd plus higher were so far down in the noise that I would hesitate to specify variation.
I can't disagree about using cheap tubes when possible; after all, I use sweep tubes and bargain-bin 6BG6 for my power amps! I've saved my D3a for phono stage use, where getting the noise performance I want from cheaper stuff is not a trivial exercise. Fortunately, D3a were relatively inexpensive when I bought them ($15 or so each, including shipping and duty when I bought a dozen).
Triode Dick has used that tube fore some time
http://www.triodedick.com/ceasar_2/ceasar2_schema_versterker.GIF
http://www.triodedick.com/ceasar_2/ceasar2_schema_versterker.GIF
Tweeker said:
Good thought. Grid stopper is close, but I can work in some SMD stoppers on a few more pins just in case.
Tubes4e4 said:
I just don't believe that, hence the desire to experiment a bit. It is a high Gm tube, but nowhere near a 7788 which I have worked with plenty.
At any rate, there are a number of tubes that will fit the board, so it isn't a one trick pony, so to speak.
Tubes4e4 said:
I have about a dozen, and while there is some variation, they are all pretty close. Maybe I just got lucky.
Tubes4e4 said:
I picked up mine for $10 each (sequential serial #'s), so this does seem to be the trend.
Tubes4e4 said:
My purpose here is driving a choke phase splitter followed by a pair of 2A3's, so low rp and high current are essential. A 6SN7 is just not going to get it done. If the D3a doesn't work out, then maybe a 5687 will do, but then I need more voltage gain. The convenience of a single tube makes playing a bit worth the effort.
At any rate, if it is a big failure of a project in the end, it wouldn't be my first.
Re: I am interested
This is really just an experimental board for myself. If a few other people want them, then I am happy to do a slightly larger run. But, I don't think this will be a big group buy item, so price is likely to stay highish. If we get people in for a dozen or so, the price would drop into the $20 to $25 range for a pair -- and might include silk screen, too.
That said, I am not sure this is the best option for a production piece. It is really meant for experiments to determine what is ideal in your situation, doing comparisons, etc. from which you would then presumably build PTP. However, I am happy to include you if you want.
If anyone wants some, send me an email ( sockets@ecp.cc ) by Wednesday and I'll look to getting an order placed around next weekend.
pred said:
This is really just an experimental board for myself. If a few other people want them, then I am happy to do a slightly larger run. But, I don't think this will be a big group buy item, so price is likely to stay highish. If we get people in for a dozen or so, the price would drop into the $20 to $25 range for a pair -- and might include silk screen, too.
That said, I am not sure this is the best option for a production piece. It is really meant for experiments to determine what is ideal in your situation, doing comparisons, etc. from which you would then presumably build PTP. However, I am happy to include you if you want.
If anyone wants some, send me an email ( sockets@ecp.cc ) by Wednesday and I'll look to getting an order placed around next weekend.
I recently just build Pete Millet's E-linear KT88 SE amp. It uses a D3A in pentode mode for a driver. I have a dozen or so D3A's that I bought off classic_components on ebay in Germany.
The tubes are all NOS Siemens. I pulled two out and popped them in when I was testing voltages. The biases were way off (>5% iirc). I was hunting down the voltages elsewhere thinking it might have been something else. Then I grabbed another tube out and replaced tube #2 with this tube #3. Still off by same amount, but realized the channels hadn't changed. So I put #2 in for #1 and voila! The voltages/bias were spot on (<1%).
I haven't gone through the collection, since I got a good match, but one tube right out of the box was way off the other pair.
Just adding that to the experience list.
The tubes are all NOS Siemens. I pulled two out and popped them in when I was testing voltages. The biases were way off (>5% iirc). I was hunting down the voltages elsewhere thinking it might have been something else. Then I grabbed another tube out and replaced tube #2 with this tube #3. Still off by same amount, but realized the channels hadn't changed. So I put #2 in for #1 and voila! The voltages/bias were spot on (<1%).
I haven't gone through the collection, since I got a good match, but one tube right out of the box was way off the other pair.
Just adding that to the experience list.
I get them off of old computer motherboards and plug in cards. We needed a bunch at work and our parts guy got them from Newark Electronics.
UH, I only got two, maybe one will work.
I have collected 2 or 3 each of several high GM pentodes, varying from the "cheap TV tube" variety to the 6688 and the D3A. I have been meaning to test them all in the same circuit for a few years. The fact that someone else makes the board, makes that job easier. I want to see if the D3A is that much better than the others in some real world circuits, or if it really is too hard to tame.
I worked in a TV repair shop in 1970 and 1971. TV's had just started using the new 6EJ7 and 6EH7 in their IF strips. Those tubes originated in Europe and most of the versions available here were relabled European tubes. It became obvious that the TV manufacturers hadn't learned to "tame" those either. Some TV's (notably Zeniths) would oscillate wildly the first time the IF tubes were changed. The later American production 6EJ7's sisn't oscillate as bad, probably because they were just crummier tubes.
There are a couple of people who wanted these boards. I am going to go ahead with a small run, so if anyone else wants some, let me know.
They will include silk screen/solder mask, and I'll put up a web page with directions on where jumpers go, reasonable parts to start with, etc.
Cost is $25 for pair, and boards should arrive next week sometime.
They will include silk screen/solder mask, and I'll put up a web page with directions on where jumpers go, reasonable parts to start with, etc.
Cost is $25 for pair, and boards should arrive next week sometime.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.