Hello everyone! As my third tube amplifier I want to be more than 3.5 watts but to use as previous one DHT as final tubes. I scouted the internet and I reached a SET schema with 4 2a3 triodes proposed by Andrea Ciuffoli. My question is how much can be the differences of Ia and GM between tubes? The topology with the cathode of 2 tubes connected together help somehow?
Thank you!
Thank you!
Attachments
Each PSE requiring tight tolerance between paralleled power tubes.
At chosen operating point (at same grid voltage Ua, Ia has tight tolerance) the Gm difference must be also as little as possible.
If there greater discrepancy, the distortion will growing.
Practical experience:
In case of manually selected -used- tubes only 1-2 pairs available from 4-5 dozens.
At chosen operating point (at same grid voltage Ua, Ia has tight tolerance) the Gm difference must be also as little as possible.
If there greater discrepancy, the distortion will growing.
Practical experience:
In case of manually selected -used- tubes only 1-2 pairs available from 4-5 dozens.
Keep the cathodes of the 2A3's separate. Each its own Rk (880 ohm - 1k ) and a 100 ohms pot in series with the Rk of the tube that draws the most current then equalize the two. Don't worry too much about inequality of the tubes; I've used this scheme more than once and it always worked out well.
I just found 4 used tubes and the values on TV7 tube tester are 56, 56, 57 and 58 with a minimum of 38 that mean almost 90% from a new tube? It is ok 90%?
Only when the filaments of the tubes have their own 2,5 v source you can adjust their anode currents separately making one of the Rk's variable. If they would share one Rk you can't because the grids are connected.
And why this would be so ? I was always under impression that matching in parallel SE circuits is not necessary unlike in push pull circuit. you could actually use two different type of tubes. Audio Matiere used that technique in Push Pull amps paralleling KT88 and EL34 in one leg of PP circuit.
But all triodes have a 2,5v filaments separately and even the substitution for 100 ohm pot, only the resistor for cathode are common for L and R channels
About a decade ago I tested some 300B tubes:
320V, 19mA CCS, 5V R.C. regulator, 1V input.
--------------
EH300Bgg_56mA/V (-68.9V, 19mA, 320V, gm:2760) THD:0.0216%
EH300Bgg_58mA/V (-68.6V, 19mA, 320V, gm:2920) THD:0.0146%
--------------
Is this a pair?
Manufacturer granted (in general 5% difference accepted -some ebay seller use even 10%-).
Measured numbers are close each other.
Tube testers has similar methods to decide, that DUT close enough to etalon tube or not.
"56, 56, 57 and 58 with a minimum of 38" shows, that tubes not weak, and IN THIS operating point not too far each other. These numbers shows nothing more.
Is these tubes appropriate for preamp (for example 19mA anode current)? Sure.
Is paralleling these 300B tubes possible? Maybe.
Is these tubes appropriate for PSE amp?
Not clearly certain with these -measured- parameters.
Tube testers (and some manufacturer too) measured tube ONLY one operating point (and with small signal), which is mostly too far for desired "live" operating point.
Some manufacturer append printed curves, but most of them "general" curves for this type of tube.
If you want to make sure that tube exactly suitable for power amp, must to measure it with "live" parameters.
-------------------
EH300Bgg_56mA/V_60mA (-70.7V, 60mA, 350V) THD:0.259%
EH300Bgg_58mA/V_60mA (-71.2V, 58.6mA, 350V) THD:0.126%
-------------------
It's the minimum requirement for the verdict: "real" Ua, Ia, bias. The THD measuring was made with 1V small signal.
Is this a pair? More or less. The small signal distortion not too bad, for individual channels can be good.
Are these tubes suitable for PSE?
I wouldn't do it.
The distortion suggests, that "first" tube grid structure not exactly manufactured well, so expected to more growing distortion with growing grid swing than case of "second" tube.
---------------------------------
So, these speculations based on some simple measurements.
If you use randomly selected, unmeasured tubes, the result is unpredictable.
Parallel operation of triode tubes can give very good results.
But in order to get very good results, it requires close attention to details.
If any of you can get permission for a reprint, you should get your hands on the very last issue of Glass Audio:
Glass Audio
Volume 12 Number 5 2000
"Paralleling Tubes Effects" (the cover article)
Kamna, McDonald, & Boehlke
600 Measurements of 20 dual triode tubes, with collation and analysis of that data; listening tests at 3 different venues using 3 different loudspeaker pair models; math proofs; and a formal proof using deductive reasoning from examples of many favorite triode tube types.
A presentation of the essence of this article was also presented at one of the VSAC conferences in Silverdale, Washington.
(Vacuum Tube Audio Conference)
Parallel operation if done Improperly has all the Bad characteristics that audiophile fans claim.
. . . Whereas, Parallel operation that is done Properly has all the Good advantages of such operation, and works extremely well.
But in order to get very good results, it requires close attention to details.
If any of you can get permission for a reprint, you should get your hands on the very last issue of Glass Audio:
Glass Audio
Volume 12 Number 5 2000
"Paralleling Tubes Effects" (the cover article)
Kamna, McDonald, & Boehlke
600 Measurements of 20 dual triode tubes, with collation and analysis of that data; listening tests at 3 different venues using 3 different loudspeaker pair models; math proofs; and a formal proof using deductive reasoning from examples of many favorite triode tube types.
A presentation of the essence of this article was also presented at one of the VSAC conferences in Silverdale, Washington.
(Vacuum Tube Audio Conference)
Parallel operation if done Improperly has all the Bad characteristics that audiophile fans claim.
. . . Whereas, Parallel operation that is done Properly has all the Good advantages of such operation, and works extremely well.
Last edited:
They are usable but not really at 90%. New 2A3's (double plate) typically test between 70 and 76 on calibrated TV7. Recent "A3's with single plate similar to 300B's do not necessarily give the same results...
Better to try with the 2a3 you have and built the amplifier.
listen.
After you could look at pairing.
listen.
After you could look at pairing.
Yesterday I experienced a disappointment because initial I hoped that will reach to have 7 watts output power, but speaking with mr Ciuffoli and studying Tung-sol technical paper the output power for OT load of 5k is 5 watts.
There is any possibility to increase output power by the topology of preamp?
Related to fixed bias versus automatic one (at final tubes) can someone to clear me up which are advantages and disadvantages of each one related to life life span of the tube, quality of sound, output power..... at the same values of operating points?
There is any possibility to increase output power by the topology of preamp?
Related to fixed bias versus automatic one (at final tubes) can someone to clear me up which are advantages and disadvantages of each one related to life life span of the tube, quality of sound, output power..... at the same values of operating points?
7W is unreal expectation.
ps. If you use it in PSE, the load would be 1k..2k5.
Up to A1, the power would be 4.5 .. 5W with 2k5 load.
With 1k65 load the power about 6W, over 6% distortion.
Lower load, higher distortion.
Last edited:
As euro21 says, 5kΩ OT is no use for 2A3-PSE.
If you want to get 7W from your 5kΩ Output transformer, change the DHT to 300B. Then you do not need parallel tubes, one 300B is enough. But the 300B must be high quality (example: EML300B), so that you can run it at 400-425V, 60mA. In this case you might get 8 - 9W, if the OT is good, and the anode supply is reasonably low impedance. Lower grade 300Bs, (like the EH300B) do not work well at high voltage (>370V) and have a short life. The EML is expensive, but they are carefully made, and if the amplifier is well-designed, they last for many years.
If you want to get 7W from your 5kΩ Output transformer, change the DHT to 300B. Then you do not need parallel tubes, one 300B is enough. But the 300B must be high quality (example: EML300B), so that you can run it at 400-425V, 60mA. In this case you might get 8 - 9W, if the OT is good, and the anode supply is reasonably low impedance. Lower grade 300Bs, (like the EH300B) do not work well at high voltage (>370V) and have a short life. The EML is expensive, but they are carefully made, and if the amplifier is well-designed, they last for many years.
Khmmm....
D3a CCS loaded, source follower, fixed biased EH300B at 400V, 70mA, 5k OPT. 4-5 Hour operation daily.
I changed one set of EH300B (gold grid) tubes after ten+ years (about 15000 hours?), because one of my tubes filament worn out. The survived tube condition was about 80%.
EML 300B it is at least 800$ a pair but for 4 6B4G I paid 180$. On the other hand 6B4G it is made in golden era of tubes in which some standard was followed, with new production of 300b we can not be sure if they have same curves like WE 300B
The quality of the EH, Sovtek (all Saratov factory) was (and probably is) very variable.
Most of the ones I tried failed with the end of filament detached from the terminal-attach point. Sometimes they arrived like this in the post. Others lasted a few hunded hours. Only one is still working - a 1998 Sovtek, that has had many years of use. It was probably assembled by the same person that made yours, Bela!
Most of the ones I tried failed with the end of filament detached from the terminal-attach point. Sometimes they arrived like this in the post. Others lasted a few hunded hours. Only one is still working - a 1998 Sovtek, that has had many years of use. It was probably assembled by the same person that made yours, Bela!