9 Cylinder amp is weak - any advice is welcome

[kagliostro]

The 9 pin tube version of the 7 cylinder amp that has build my friend is weak, only 3.12W and too much clean

In = 200mV pp >> out ef86 = 7,4V pp >> TS + Vol >> in PI = 2V pp >> out PI = 30V pp >> output on 8ohm load = 14V pp = 3,12W

here the schematic with voltages



I've think to add a Source Follower between the ef86 and TS as to recovery the lost gain

then to use the SS rectify as to have a higher B+

on the datasheet of the 6v6 for PP AB1 operation I've find this data

Va=250  Vg2=250  Vg1=-15.0  Ia=70.0(the sum of the two tubes)  Ig2=5.0  Zout=10,000    Pout=10.0   

as the 7 cylinder amp uses 6aq5 that are the same of 6v6 we supposed that also 6v6 can work correctly with the same PT (450v CT) and OT (6.6k) 

but seems we were wrong

may be the EZ81 is also a problem, because has a drop of voltage higher than the 6X4 that is used in the 7 cylinder ?

any idea on how to have the amp working correctly ?

Many Thanks

K

[shooter]

Hey, looking at the 250vdc AB1,  if my math is close, 13vdc(Vk)/220ohms = 60mA Ia, minus what G2 uses, datasheet says 70mA.  the datasheet I'm looking at says 30vac peak grid to grid.  So your ideas should get you there, either boost current, maybe 180ohm cathode, raise B+ with SS, and get more drive to G1.  I try....to get datasheet values at G1 with gain and volume knobs set on "7" and TS knobs set at "5"  that gives the player enough knob to overdrive, but not enough to squeal like a stuck pig.

[eleventeen]

250 B+ could definitely be a problem. Is that enough for a guitar amp? Maybe for a record player.


A source ("cathode") follower always has gain < 1 so I do not think that will help.


I think I would try SS rectifier, just pull out the EZ81 and tack in 2 diodes. Very easy.
 

[kagliostro]

Thanks Shooter & Eleventeen

@ Eleventeen

You are right the gain of a Source Follower is max 1, but the purpose of the Source Follower isn't to add gain

is to maintain the gain of V1 that reach the PI at a higher level

usually you don't see TS after pentodes because the TS load too much on a pentode, so a Source Follower acting as an

impedance adapter, save the gain that the pentode produce without an excessive waste on the TS

For this reason I've think to use one

Franco

[eleventeen]

Do you think your B+ @ 250 volts is enough? That seems low.

[HotBluePlates]

... only 3.12W and too much clean

In = 200mV pp >> out ef86 = 7,4V pp >> TS + Vol >> in PI = 2V pp >> out PI = 30V pp >> output on 8ohm load = 14V pp = 3,12W

How were the measurements made?

Your use of volts peak-to-peak suggests you had an oscilloscope for voltage measurement. Did you also use a sine wave generator? Did you turn up the sine wave until you heard some distortion in the speaker?

If you didn't use a sine wave signal generator (maybe just a guitar or a CD player to make sound), there's no safe assumption I can make about converting peak to average or RMS signal level. If you didn't turn up until you heard distortion, then back off volume until what you consider "maximum clean power," then I can't make a safe assumption relating measured speaker voltage (and the output power implied) to the maximum output power of the amp.

I'd suggest starting with considering the output stage first and ignoring the preamp until it is known-good, because 3w falls well below the 10w claimed by the 6V6 data sheet with 250v plate & screen and a 10kΩ OT primary impedance. But before jumping to conclusions about the OT, I wanted to verify the numbers.

The ideal way to measure peak-to-peak output voltage from the phase inverter would be to connect meter or oscilloscope leads from one 6V6 grid to the other. Your maximum clean output power will probably occur near where the peak 6V6 grid voltage (or half the peak-to-peak number) equals the bias voltage.

Once you have that input drive, and are hearing the onset of distortion, measure voltage across the speaker or dummy load. we can work backwards to figure out what the output section is doing from there.

Afterwards, you can deal with whether the preamp causes the amp to stay too-clean, too-high on the volume knob.

[kagliostro]

My friend used the oscilloscope and I think he used a signal generator, he is radar technician and has access to good equipment, but he lives

in Neaples, very far from here and I don't know the other details you are asking for, sorry (but I can ask if you want to know something of specific)

I can say he tried to use SS rectify and he had 312v at the first node of the filter, without the tubes inserted and 265v with tubes inserted

the drop seems a bit high to me, but 265v seems fine for the 6v6

must be considered that the OT is an Hammond 1650H with 6.600ohm primary and not 10k primary

thing that I think has an influence (or not ?)

Franco

[HotBluePlates]

must be considered that the OT is an Hammond 1650H with 6.600ohm primary and not 10k primary

thing that I think has an influence (or not ?)

The first thing to know is whether the measurement at the speaker terminals is for maximum power at the onset of distortion, so that we know whether the impression the amp is weak is about an output stage issue.

Let's say he measured output power well below the amp's full power capability. That could lead to an impression of "weak output" and "too clean." Then the answer could be as simple as "drive the amp harder."

Or let's say the amp really is only making 3w at the onset of distortion. Then I'd spend time drawing a loadline for the OT used and compare it to a loadline for a 10kΩ OT. I'd also work backwards by finding the turns ratio of the OT, convert the speaker voltage measured into the corresponding primary voltage swing, and see if the measurements show the OT just isn't the best impedance for the existing supply voltage.

I'm not so much worried by preamp issues. You can always find ways to add gain.

[kagliostro]

Thanks HotBluePlates

I'll ask to my friend to join us here

Franco

[jazbo8]

If the PI is really putting out 30Vpp, that means each output tube sees 15V, so that's enough to drive it to clipping if the bias is indeed 13V as shown on the scheamtic. In which case, you should get around 6W, not 3W. Did you verify that the speaker is 8R and is hook up to the 8R tap on the OPT? I ask this because, 4R on the 8R tap would get you about 3.5W.

[kagliostro]

good clue on which investigate


Thanks Jazbo8


Franco

[raf71]

Hi to all.
I'm Francos friend......excuse me for my terrible english.

About the measurements, i've a signal generator and an oscilloscope.
I've drived the preamp to have a not distorted signal. My idea was to check with 200mVpp the output power on a 8ohm dummy load, and after, increase the input signal until the output distortion.
The TO is an hammond that have 6600ohm primary Z and 8ohm secondary Z.
I've tried also to remove the EZ81, but i 've only 265V, if i remove the 6v6s the voltage, on the first cap, is 310V. If I connect in series the two dummy load that i have (8ohm + 4ohm) the output power arrive to 5W
Now i'm at work, but in the evening i will make the misurement that HotBluePlates (thank You) have suggest.

Excuse me for my english.
Raffaele

[raf71]

Hi to all.
I've perform other measure.
On g1 of 6v6s the voltage is 35Vpp undistorted.
On the anodes the output voltage is 250Vpp
On the dummy load (8.2ohm) the voltage is 15Vpp
The screen current is 2.3mA
The voltage on the G1 is -15,5V
The katode current is 35,8mA

The amp dont has the feedback, is possible that i must be invert the anode wire? but i dont see any oscillation on the output.

Best regard to all.
Raffaele

[kagliostro]

May be I'm wrong, I'm not very much confident with math, but ...

15Vpp on a 8.2R load isn't 27.43W ? (15x15 = 225 -- 225/8.2= 27.43W)

assuming 1/2 of 27.43W we have 13.7W that isn't so bad for a pair of 6v6

Franco

[jazbo8]

No, 15Vp-p = 5.3Vrms, Po = 5.3² / 8 = 3.52W.

[kagliostro]

OK, thanks Jazbo8

and 3.52W from a pair of 6v6 is really low


K

[jazbo8]

Did you guys find out why the output was so low? Enquiring minds want to know...

[DummyLoad]

Vrms = 0.3535 * Vpp


--pete

[kagliostro]

Vrms = 0.3535 * Vpp

Thanks Pete

Did you guys find out why the output was so low?

not yet 

Franco

[drgonzonm]

Have you tried replacing the 12AY7 with a 12au7 or a 12ax7 to see if that makes a difference? 

This is a long shot, and I ask other forum members to critic, try removing the capacitor on the PI, to increase gain. 

[HotBluePlates]

Franco,

Have your friend place a 16Ω load on that 8Ω secondary, drive the output tube grids to the same level (30v peak-to-peak, measured from grid to grid), and re-measure the voltage across the load.

I'm betting it goes up, indicating more output power with no circuit changes.

[kagliostro]

My friend has other family commitments this days, he got one other PT to be tried, with higher voltage

He has controlled the OT connecting 6.45v on the secondary and he has 162v on the primary (81v+81v)

so the transformer ratio is 630,83 from which he calculated 5046.64ohm impedance with a load resistor of 8.2ohm (at 50Hz)

may be a higher frequency voltage give the labeled 6.6k

without the ez81 on the circuit, and using a pair of SS diodes B+ is near 310v unloaded that decreases to 270v when loaded

and he is convinced the 75mA available  at the PT (Hammond 369GX) are not enough for the amp consumption

He has done also a test using a load of 8.2R+4R = 12.2R (he has not 2 x 8.2R power resistors) and he told the power increased at near 5W

This evening he would like to go on with the amp, but he is engaged with his daughter to go to the cinema to watch a movie (I think a cartoon)

Franco

[HotBluePlates]

The OT primary impedance is 100% set by the turns ratio between primary and secondary, which is why he can do the test to get a primary Z value; Changing the frequency will not alter that impedance. That said, at extreme low/high frequencies, other parasitic effects will reduce through-put of the OT.

Where I was going with the request is that I think a higher primary impedance would have yielded more output power; something closer to the 10kΩ impedance shown on the data sheet for 250v plate & screen. So unless he's already paid for the new power transformer, he may want to try that.

Keeping the 5-6kΩ OT and increasing B+ a lot (like up to the levels of a Deluxe Reverb), while also biasing the amp cooler at idle will increase output power. That doesn't make this the "right condition" but only that increasing B+ voltage will raise output power, while have lower primary Z (than the ~10kΩ ideal load for a 200v condition) will raise current, again increasing output power. Overall, this is going from Class A to Class AB.

Sorry I was out & unavailable for a week or two... My thinking from the start was that the drive into the output tubes grids was at least equal to the bias voltage, so the output tubes weren't being under-driven by the preamp. The abnormally-low output power then had to be an issue about having a non-optimal load for the tubes. It is often overlooked that there is not a single ideal load for a tube, but the load giving maximum output power depends on supply voltage and intended class of operation.

Either way, I'm estimating an optimum load for the amp would give ~8w RMS, much less than what you're used to for push-pull 6V6's (because of the low supply voltage and class A operation). This is more than double what your friend measured, which says the loading is wrong in his case.

[kagliostro]

Many thanks for the explanation HotBluePlates

I told my friend to try with a 16ohm speaker on the 8ohm intake as from your previous council

he also received a 300v transformer from a friend and he want to give a try with a higher voltage, and current (the PT he has now has 75mA)

The transformers he has installed on the amp at the moment are those indicated on the 7 cylinder project

PT Hammond 269GX (OK, the 369GX which is the european version)

http://www.hammondmfg.com/pdf/EDB369GX.pdf

OT Hammond 1750H (that for some reasons didn't seems to arrive to the 6.6k at which it is labeled)

http://www.hammondmfg.com/pdf/EDB1750H.pdf

--

Talking about the OT a friend is building an OT for my version (with 7 pin tubes) of this amp

I asked for a 5000 + 5000 (10k plate to plate) and the PT I've is rated for 225v AC

The most datasheet for 6v6 I've seen say B+ 250v - 10k OT - 10W output power

I've seen only one where the 6v6 with B+ @ 250v is loaded with a 12k OT and give 13W output power

http://www.mif.pg.gda.pl/homepages/frank/sheets/084/6/6V6GT.pdf

But I want to use 6aq5 tubes and I think they are better suited with 10k also for dissipation reasons, which is your opinion ?

Thanks

Franco

[HotBluePlates]

But I want to use 6aq5 tubes and I think they are better suited with 10k also for dissipation reasons, which is your opinion ?

Use whichever you prefer. The 6AQ5 is a 6V6 in a 9-pin bottle. When you look up the 6V6 in the 1959 RCA tube manual, it refers you to the 6AQ5 entry for plate curves.

[kagliostro]

Thanks


Franco

[DummyLoad]

But I want to use 6aq5 tubes and I think they are better suited with 10k also for dissipation reasons, which is your opinion ?

Use whichever you prefer. The 6AQ5 is a 6V6 in a 9-pin bottle. When you look up the 6V6 in the 1959 RCA tube manual, it refers you to the 6AQ5 entry for plate curves.

Respectfully, 6aq5 is 7pin 6v6. 6cm6 is 9pin 6v6.

--pete

[kagliostro]

Thanks Pete

That is an interesting info

I think the 6cm6 are similar to the 6P1P-EV I had from an hungarian radio amateur at the Ham Fest on Marzaglia in 2014 at 3.4 $ each

http://www.mif.pg.gda.pl/homepages/frank/sheets/127/6/6CM6.pdf

http://www.mif.pg.gda.pl/homepages/frank/sheets/113/6/6P1PEV.pdf

Franco

[HotBluePlates]

Respectfully, 6aq5 is 7pin 6v6. 6cm6 is 9pin 6v6.

Thanks! I usually get myself in trouble when talking bout tubes I don't use & not looking them up, and this time was no exception!