Again OT: push-pull 30V tube amp output transformer?(sopht)

[Steben]

Hi there guys.

I make an attempt again to get info about making a push-pull version of the sopht amp(s), achieving maybe 1 Watt.
What output transformer would I need? Single-ended, the 12K5 needs 600ohm. P-P 1200ohm?
The only thing is see is the hammond 125A then, with lowest impedance.

[slajeune]

Hi Steben,

yes, the 125A with a 4ohm speaker will get you close enough.  Apparently you can also use a 70V line transformer as a push pull transformer, I just don't have the details on it.  Finding the bias for the 12k5 will be the hardest part.

70V info:

Ok, I got the required info.  I haven't tried this but from my readings, use the 2.5W as the center tap and use the common and 0.62w as the plate to plate winding.  With a 4ohm speaker load, the plate to plate will be aroun 1200ohm which is right on for the 12k5.

You could also investigate the 26A7-GT which is a tube that is meant to run at 26.5V (heaters and plate) and will put out about .5watt in push pull.

Hope this helps!

Thanks,
Stephane.

[iccaros]

This is old, but I need math help

I want to put two 12k5 in P-P and I have a 70volt line transformer @ 10watts. it has .62, 1.25,2.5, 5, 10w and a common.

so what is the split.. I see from Slajeune that they say 2.5 and .62 but how did they come up with that?

the math I think is right...
70v so...  70 sqr / watts
70*70/2.5 =  1960ohm.. Now speaker means nothing as the transformer has 8 and 4 ohm taps.. which tells me that @ 4 or 8 you will have the same load as long as you use the correct taps..

Now if you put 8 ohms on the 4 ohm tap you should get double the load.. 1960*2 = 3920...


Am I doing this right.. or an I way off..

[amptramp]

I also use matching transformers for outputs - large numbers of them were used for discos, malls and other such installations where large numbers of speakers were used, so they are plentiful and cheap.  If you have a standard 70.7 volt transformer, power and resistance are related as follows:

P = I * V

I = V/R

Therefore

P = V²/R and R = V²/P

In this system, resistance is actually the impedance and is not related to the winding resistance.  DC resistance will not correspond to the impedances since the windings closest to the core will be lower resistance and the windings near the outer side will be higher in resistance due to the greater circumference, so do not look for matching DC resistances.  In this case, V = 70.7 so V² = 5000 and the impedance can be derived from the power indicated for that tap.  If you are making a push-pull transformer from a matching transformer, you want the impedance to be four times as high because doubling the number of turns on an inductor gives you four times the inductance (which is most of the impedance).  Using the formula above, 2.5 watts is a 2K impedance and you are therefore looking for an 8K impedance to give you a balanced push-pull with the same number of turns on each side, from the 2.5 tap to the common and the other tap you select.  Therefore, you would need 8K across the entire winding and that is the 0.625 watt tap.  Using a 4 ohm speaker on the 8 ohm output cuts all the effective impedances in half, so you end up with 1K each side.  This is close enough to the 800 ohms specified for the 12K5.  (Now maybe someone can chack my math and make sure I haven't screwed up somewhere.)

[iccaros]

I also use matching transformers for outputs - large numbers of them were used for discos, malls and other such installations where large numbers of speakers were used, so they are plentiful and cheap.  If you have a standard 70.7 volt transformer, power and resistance are related as follows:

P = I * V

I = V/R

Therefore

P = V²/R and R = V²/P

In this system, resistance is actually the impedance and is not related to the winding resistance.  DC resistance will not correspond to the impedances since the windings closest to the core will be lower resistance and the windings near the outer side will be higher in resistance due to the greater circumference, so do not look for matching DC resistances.  In this case, V = 70.7 so V² = 5000 and the impedance can be derived from the power indicated for that tap.  If you are making a push-pull transformer from a matching transformer, you want the impedance to be four times as high because doubling the number of turns on an inductor gives you four times the inductance (which is most of the impedance).  Using the formula above, 2.5 watts is a 2K impedance and you are therefore looking for an 8K impedance to give you a balanced push-pull with the same number of turns on each side, from the 2.5 tap to the common and the other tap you select.  Therefore, you would need 8K across the entire winding and that is the 0.625 watt tap.  Using a 4 ohm speaker on the 8 ohm output cuts all the effective impedances in half, so you end up with 1K each side.  This is close enough to the 800 ohms specified for the 12K5.  (Now maybe someone can chack my math and make sure I haven't screwed up somewhere.)

so the good point is my math was right (I don't normally round up but I noticed most do) , I was just missing the other side, which is number of turns, so instead of doubling the number I double the windings giving me 4x the impedance.  or to think about it my way.. I have 2K from one side (0+ 2K) so I need 2K from the other to split it (10K - 2K = 8K) which gives me the same  number of windings from the common which is center?

[amptramp]

The common is not the centre tap.  The common goes to one of the plates.  The 2.5 W tap is the centre tap and the 0.625W is the other plate.

I don't know where the 10K - 2K is coming from in your reply.  You need some number of turns from one side to the centre tap and you need the same number of turns on the other side.  If you double the number of turns, you have four times the inductance which is also four times the impedance - that's why you go to a tap that is 1/4 of the power of the one you are using as the centre tap.

[iccaros]

sorry that should have been watts. Its is a 10watt 70 volt transformer, so I was think out loud ... wrong..  I think I have it.. I am not that quick

[bancika]

You could try using a power transformer for that. 230V:18V tranny with CT-ed primary has ratio of 12.7, which makes it suitable as a 1.3K:8ohm output tranny...

[iccaros]

I could, but I already have 4 line transformers.. I can only spend $60 a month.. So with tubes this month, I blew that

[iccaros]

@Amptramp

If I could ask one more question...

I have a 26A7-GT that I want to use in place of the 12k5.. This tube is a dual pentode..  http://www.mif.pg.gda.pl/homepages/frank/sheets/049/2/26A7GT.pdf
it states effective loading plate to plate is 2500. So do I need 5K for this push pull to work?

In their example they show a transformer stating 2000 plate to plate load (that is not a 2000 CT is it?) .. If its 4000 I am good, I have a Hammond rated at 2000 or 4000 depending on what taps and speaker.. SO I could also get up to 8000 using a 8ohm on the 4 ohm tap or as low as 1 K using 4 ohm on the 8 ohm tap...


Thanks

@bancika - sorry, my statement did not come out right.. I appreciated the option.. and I may look for some of those @ good will when I go, but I have the 70 v line transformers.. So they were my first option..
I have seen people using power transformers.. I am more likely to do the opposite.. Use the power transformer backwards to get high voltage from a wall wart..
But thanks..

[amptramp]

For the 26A7, your source shows a load resistance of 1500 ohms individually or a plate-to-plate resistance of 2500 ohms.  Using your line transformer with 0.625, 1.25, 2.5, 5 and 10 watt taps, you would use common as one plate, 5 watt as the centre tap and 1.25 watt as the other plate.  This gives you 4000 ohms plate-to-plate which would not be the most efficient coupling, but it is safer to use a higher impedance than to go one step lower and use 2000 ohms plate-to-plate.  The higher impedance also gives better bass response because the inductances are higher.

With the Hammond 125A to 125E transformer, you have the option of lots of plate-to-plate impedances.  Make sure you have the push-pull and not the single-ended transformer.  You are limited to 2 watts per section with this tube, so anything from 2000 ohm to 3000 ohm is usable with the higher values being a safer choice for limiting tube current.  Even though the Hammond impedances are set from the secondary as per:

http://www.hammondmfg.com/125.htm

and the primary inductance is fixed, the loads reflected back to the primary set the impedance.  It is not that critical - I have a matching transformer mounted in a box with a selector switch for testing my antique radios and there is best power delivery at the recommended impedance, but it is not critical.  Note that your speaker will not necessarily be at the rated impedance over the entire frequency range, so some experimentation may be in order.  If you use the 125 H or 125J series, the secondaries are tapped but with fewer choices and the 125J provides a 2500 ohm plate-to-plate impedance.

[Johan]

a small power toroid transformer, 230V CT (115+115volt)primary and 24volt (2x12v) secondary could do your work at 8 or 16 Ohm
J

[iccaros]

For the 26A7, your source shows a load resistance of 1500 ohms individually or a plate-to-plate resistance of 2500 ohms.  Using your line transformer with 0.625, 1.25, 2.5, 5 and 10 watt taps, you would use common as one plate, 5 watt as the centre tap and 1.25 watt as the other plate.  This gives you 4000 ohms plate-to-plate which would not be the most efficient coupling, but it is safer to use a higher impedance than to go one step lower and use 2000 ohms plate-to-plate.  The higher impedance also gives better bass response because the inductances are higher.

With the Hammond 125A to 125E transformer, you have the option of lots of plate-to-plate impedances.  Make sure you have the push-pull and not the single-ended transformer.  You are limited to 2 watts per section with this tube, so anything from 2000 ohm to 3000 ohm is usable with the higher values being a safer choice for limiting tube current.  Even though the Hammond impedances are set from the secondary as per:

http://www.hammondmfg.com/125.htm

and the primary inductance is fixed, the loads reflected back to the primary set the impedance.  It is not that critical - I have a matching transformer mounted in a box with a selector switch for testing my antique radios and there is best power delivery at the recommended impedance, but it is not critical.  Note that your speaker will not necessarily be at the rated impedance over the entire frequency range, so some experimentation may be in order.  If you use the 125 H or 125J series, the secondaries are tapped but with fewer choices and the 125J provides a 2500 ohm plate-to-plate impedance.

I have a 125J, I got just for this, I wanted to make sure I picked it right though..
On the last page they have a example design showing 2000 ohm plate to plate, but in the data sheet it says 2500, so I was not sure.. But I understand, If I use higher its ok, may get less power but better sound.

And last.. you have explained this better than any one else I have talked to.. If a tube does not list a load resistance, how do I figure that out.. I have been told to divide by 6 the plate resistance, but that made no sense..

Thanks for the help,  anytime I can pay it back I will..

[PRR]

> they say 2.5 and .62 but how did they come up with that?

Actually, for tapped-primary line transformer, it is:

Common
quarter-power
lowest power

On a 10 Watt transformer, the 2.5W tap is the center. (Half the number of turns.)

The lowest-power (all the turns) may be 1W, 0.6W, something like that. As long as it is MUCH lower than the quarter-power tap rated watts, it's close-enough to work.

The 12K5 is NOT fussy about the exact impedance. (It works bad for any impedance; 400 to 1600 per plate is the least-bad zone, in flea-watt amps 100 to 5K per plate may all give small useful output.)

> 26A7- dual pentode.. 

TWIN pentode. ("Dual" can be two different pentodes.)

> it states effective loading plate to plate is 2500.

PLATE TO PLATE means the same as "2.5K CT".

26A7 is a hair more fussy than the soft-knee 12K5 but 2:1 variations cause little power change. If you had to meet specific goal for power and THD into a known load, you'd want to get close. In terms of a small music amp driving a loudspeaker, you don't have to get real close.

> I have a Hammond rated at 2000 or 4000

If that is 2000 CT, then you are near-enuff spot-on the 2500CT suggestion for AB1 Push-Pull. 4000CT will give a shade less watts with somewhat different distortion flavor. Try both.

In a cathode-biased tube amp, "too low" load impedance will NOT hurt the amp. (Tubes are not transistors.) The amp is Class A, which means it takes MAX power at idle. You can only get out what it sucks in. If you do take large power out, the tube(s) run cooler. But in speech/music audio duty, the tube MUST survive "idle" for long periods. And cathode-biased amps will "block" and tend cooler if over-driven (under-loaded). All you can do is get less power and more flavor. Since you are rejecting the "big" tubes like 6V6/EL84/6L6/EL34, obviously "less power" is not a problem.

BTW: self-bias tube amp speaker jacks are often wired to SHORT the amp when speaker is gone. This is safer than leaving the OT unloaded, flyback kicks.

(FIXed-bias tube amps can be over-strained with too-low impedance loads, same as naked transistor amps.)

-----------

> If a tube does not list a load resistance, how do I figure that out..

It is a loooooong lesson.

Ya know, nobody EVER asks what load impedance to use with a specific transistor (NPN, FET, IGT, etc). Why not? (And we -did- watch our matching when transistors were new.)

> I have been told to divide by 6 the plate resistance, but that made no sense..

No and yes.

If a tube was SOLD as a power tube, it will have a Suggested Load on the sheet.

This will very-very often be around 5X to 10X the pentode plate resistance. "6X" is not wrong (but is more specific than necessary).

Note that pentode plate resistance is high, uncertain, and varies a lot with current. This parameter is not carved in glass.

For PENTODE (including so-called power tetrodes), there are TWO plate resistances. Above the knee and below the knee. For 26A7, about 5K above the knee and about 150 ohms below the knee. The good-power load is well-away from either number. For one tube, below 2K but above 300 ohms. The best-power load may be the geometric mean, about 866 ohms. But it is more important to stay away from the two plate resistances than to be right-on the geometric mean.

Note that many power pentodes don't cite the higher plate resistance, and no tube sheet says what the lower plate resistance is (not even as purveyance).

I will say, without proof here, that if you can find or estimate the triode-connected plate resistance of a pentode, then 2X to 5X that impedance will be a good load.

And that most pentodes will work well in that 2X-5X load if the ratio of plate voltage and plate current is adjusted similar to load impedance and plate dissipation is raised as high as you wish/dare (respecting all other rated limits).

The suggested load impedance may be changed over a wide range, IF you vary voltage and current accordingly. 6550 SE may be worked at 1.7K with 250V 150mA, at 2.5K with 325V 130mA, at 4K with 400V 100mA, at 5K with 455V 91mA, or at 10K with 650V 65mA. There's not a big difference in power or efficiency over this range. Note that in each case, V/I is about Rl, and V*I is about 40 Watts idle dissipation (tube limit). (If you read the 6550 sheet, they suggest somewhat lower loads... this fine tune must be found by tedious experiment, and may give great numbers on the sheet, but is meaningless for loudspeaker use.)

[iccaros]

Thanks
PRR
and AmpTramp

and all others..

So in processing.. One thing that struck me, is not the OT acting like the Plate Resistor for the output tube, so I would use it to draw the load line.. ensuring that what I pick dose not cause the tube to draw more current than the rated max dissipation. This is along with the cathode bias...

I found with the 12k5 that under 980 ohms.. It just can not be clean.. And @ 980 clean is for a very short period.. But preamp has something to do with that.  but its a fun tube, you can change settings on the fly...
As for when no speaker I have a 20watt 8ohm resistor on the pins when not plugged in.. This is a simple thing to add with the jacks that break contact when you plug in..