strange output transformers

[dscottguitars]

Hi, I got this old 3 channel hammond organ amp that has two of the PP output transformers strangely wired.  They have a blue and yellow wire going each plate.  Red wire to HV, brown and green wires going to the speaker.  This is typical and I've used this before.  But, there is also an orange and white wire going to each cathode and a black wire going to a capacitor-25uF/25VDC.

Has anyone seen this before?  I was going to use this for an 18 watt amp similar to your Stout schematic, but I'm not sure about the wires going to the cathodes and the capacitor.  Can I just leave them unconnected and still use the transformer?  I've already drilled for its size...

Thanks,

Daniel

[sluckey]

But, there is also an orange and white wire going to each cathode and a black wire going to a capacitor-25uF/25VDC.

Might be a negative feedback circuit. Post a schematic or model number and we may be able to be more specific.

[dscottguitars]

I don't have a schematic but the model is PR-20.

On the chassis is H-AO-33-4. 

Transformer numbers:
H-AO 21264-1 stamped on the housing
AO=21280-1 on the paper inside the amp.

It has a separate power supply too...

Thanks...

[dscottguitars]

At first glance I thought they were ultralinear, but did not look closely at where the wires went until I took them out...

[kagliostro]

Schematic

http://www.captain-foldback.com/Hammond_sub/schematics/pr20.zip

K

[HotBluePlates]

I don't have a schematic but the model is PR-20.

On the chassis is H-AO-33-4. ...

Always check Captain Foldback for Hammond schematics.

If you look at the schematic for your amp there, you will see a 130Ω resistor and 25uF cap attached to a winding on the secondary side of both output transformers. This winding is separate from the secondary attached to the speakers.

I agree with Sluckey: based on the schematic, this looks like a feedback winding, applying feedback to the output tube cathodes. For convenience, the shared cathode bias resistor and bypass cap are attached to the center-tap of this winding.

NOTE: Had that little winding been drawn on the primary side instead, it would have indicated a split- or distributed-load for the output tubes (like McIntosh amps). While that arrangement does provide feedback, it also indicates a different mode of operation for the output tubes.

If you don't want to wire the output stage like Hammond, just heat shrink and tuck away those 3 leads for the feedback winding.

[dscottguitars]

Thanks for that link!  That helps greatly...

Is there any advantage/disadvantage to using that circuit?  It seems like more feedback voltage, but going into the power tubes where I have put feedback into the phase inverter.  Here is the power section I've used before and plan to use here.

[HotBluePlates]

... It seems like more feedback voltage ...

Nobody knows how much feedback it is unless they measure it. That's because the winding can be any ratio in relation to the primary, and was determined on the basis of satisfying the original Hammond circuit.

By comparison, the feedback voltage from a typical Fender circuit can be calculated, because a specific voltage is implied by assuming a power output level and the load impedance.

You could always try it first and see what you think of the sound. Tying with and without feedback would be a little complicated as you'd have to re-rig the cathode resistor and bypass for the 2 different situations.

... Here is the power section I've used before and plan to use here.

If you plan on keeping EL84's, then you should use a ~130Ω resistor (possibly another value on test) rather than the Fender-style 220Ω.

That's because the 220-250Ω resistor is typical for a pair of cathode-biased 6V6's or 6L6's, but the EL84 has higher Gm and therefore uses a smaller cathode resistor and a smaller bias voltage to set its idle current.

[jojokeo]

Is there any advantage/disadvantage to using that circuit?  It seems like more feedback voltage, but going into the power tubes where I have put feedback into the phase inverter.  Here is the power section I've used before and plan to use here.

All looks about "normal" just slight changes from standard values. But the feedback resistor is pretty low giving you a high feedback ratio of 8.4 - too much IMHO. Better to start w/ around 20:1 off the 8ohm tap using 100k & 4.7k or even 100k & your 5.6k and adjust if necessary. Too much feedback will 'anchor away" your tone down.

[dscottguitars]

Thanks...

That cathode resistor value I arrived at in another amp I built using the Weber bias calculator website.  If I remember right, I had about 12v on the cathode which the 220 ohms put the current around 26mA with 325v on the plates.  I started with a 150 ohm and it seemed too high according to the calculator, but I do see most circuits with that smaller value.  Could that be due to the higher voltage I have on the screens? 

For the feedback, I'm going for a cleaner sound but I'll keep your suggestion in mind and may settle somewhere in between.

[jojokeo]

Got it but even Fender had a 15 - 17 x ratio in his Twin, Princeton, & others and those were "clean" amps. Too much takes your tone down with it and yours will half of that, just sayin' but it's your amp

[dscottguitars]

I'm not sure what you mean then, the Twin, schematic ab763 from the sixties has an 820/100 feedback ratio.  The Princeton has 2700/47.  Those are way different. 

[HotBluePlates]

That cathode resistor value I arrived at in another amp I built using the Weber bias calculator website.  If I remember right, I had about 12v on the cathode which the 220 ohms put the current around 26mA with 325v on the plates.  I started with a 150 ohm and it seemed too high according to the calculator, but I do see most circuits with that smaller value. ...

Maybe the best plan then is to do PRR's trick: Get a 50-100Ω resistor and a handful (about 15) 10Ω resistors. Put all in a series string, with the highest value connected to ground. Get a jumper wire with alligator clips on either end, clip one end to the tube cathode, and progressively short out unneeded 10Ω resistors one-by-one until you land on your desired idle current.

It's just an experience thing... 6L6's and 6V6's, with reasonable supply voltage for cathode bias wind up with a higher bias voltage and a larger cathode resistor than EL84's in the same circuit. EL84's require less bias voltage to keep them in check (like 10-12v for an EL84 vs 18-20v for a 6V6).

And it seems to be the rule that everyone guesses too low with their bias resistors to start, so it's probably better to have an adjustable rig at first until you know for certain what will work in the actual amp.

[PRR]

> Had that little winding been drawn on the primary side instead, it would have indicated a split- or distributed-load for the output tubes (like McIntosh amps). While that arrangement does provide feedback, it also indicates a different mode of operation for the output tubes.

Disagree.

There's only one core. Which side we draw the windings may aid thought, or may just be convenient for the drawing.

The Mac's cathode winding is not Positive feedback, but a (partial) Cathode-Follower. (Some models have added windings to screen, and to driver; the driver winding is bootstrap, Positive feedback.)

The key difference between this Hammond (also seen as a mod for Dyna) and Mac is that Hammond's cathode winding is maybe 10%, Mac took a full 50%. So Mac is MUCH more linear, but also has a BIG drive problem.

Modes: the self-bias Hammond is clearly run near class A. The bigger Macs run cool AB, even AB2. The huge NFB from the 50% cathode winding keeps them linear over a much wider range, so cool AB is still clean. Pushing a small grid current gains a few more Watts (the magic "50" number) within the spec-limits of the older tubes.
___________________________________

> any advantage/disadvantage to using that circuit?

Tube Hammonds are about the most "musical" amplifiers ever built. The guy had style and knack. It would be very reasonable to adopt the whole output stage (S3 onward) and play through it. (Might want larger coupling caps; in fact as a "treble" channel that OT may not like big guitar bass.)

But if you have your own known-safe circuit, just ignore that cathode winding.

[jojokeo]

I'm not sure what you mean then, the Twin, schematic ab763 from the sixties has an 820/100 feedback ratio.  The Princeton has 2700/47.  Those are way different. 

So then what's different about the two examples? The Twin's got two speakers and a different impedance tap than the single speaker Princeton. This makes a difference.

What's the tap on your transformer that you plan to use w/ how many speakers of what impedance exactly? You need to know this before ball-parking NFB ratios. It would be beneficial to you to build the amp and initially make the ratio high and get it running well. Then you can always simply lower it to exactly what you like when ready. This will save you headaches in the long run dialing in the amp properly first.

[dscottguitars]

From what I know, the number of speakers should not matter only the impedance.  I'm using 8 ohms.  Upon some research I found:

Princeton: ~15W@8ohms = ~11v/(2700/47)= 0.19v of NFB or 1.7%

Twin:  ~85W@4ohms = ~18.4v/(820/100) = 2.25V of NFB or 12.2%

That's more than 10 times the NFB voltage for the Twin.  Are my calculations correct?  I would like to know for future reference when deciding what ratios to use.

This tells me the Twin should be cleaner than the Princeton.  I usually shoot for about 15-20% for clean and less than 1% for max distortion.

Let me know...thanks

[dscottguitars]

The Princeton schematic I found at Ampwares shows the blackface a 6V6 push pull amp.  That's what I based my info on and info at Fenderguru...

[Willabe]

Fender made both a SE Princeton and a PP Princeton.

Look in the amp schematic library here and you can see them both.

The tweeds were SE, 5B2-5F2A. Then they became PP starting with the brown face 5G2, then on to the BF AA964.

http://www.el34world.com/charts/Schematics1.php


            Brad    

  

[PRR]

> Are my calculations correct?

Incomplete.

You must compare the 0.19V or 2.25V against the signal level at the point NFB is returned to.

The 15W is surely 6V6 which run ~~~20V max at their grids, the Twin is hi-volted 6L6 and may want 40V at its grids.

I'm too lazy to look up specifics. But is the 0.19V amp's driver a gain-stage plus a split-load, and the big amp uses the long-tail affair? The long-tail has half the gain.

So counting on thumbs, that's a 1:4 difference in signal level at the point NFB comes back into the amplifier.

If you really want a constant "% NFB" number.... wire the amp without NFB. Load in correct load. Bring it up with a midrange test-tone to a couple Volts. Now apply NFB and tinker. The point where output drops to half (in real use, you must hit twice as hard for same output) is a convenient benchmark. That's pretty-near where most Fenders run. 6dB NFB. If you only NFB so output drops 10%, 1dB NFB, it is pretty raw but has some damping on bass-resonance. If you apply so much NFB that output is 1/10th, 20dB NFB, you are cleaner than many-many classic hi-fis. This will probably be too low-gain, also very prone to oscillate. Numbers like 75% to 25% drop are reasonable for g-amps.
___________________________

> princeton is a Single ended

I think the Princeton badge was used on both one-6V6 and two-6V6 models? {EDIT} Confirmed while I was typing.

[HotBluePlates]

... The Mac's cathode winding is not Positive feedback, but a (partial) Cathode-Follower. ... Hammond's cathode winding is maybe 10%, Mac took a full 50%. So Mac is MUCH more linear, but also has a BIG drive problem. ...

The bad thing is this is what I meant to say; i.e., distributed loading puts half the load in the cathode circuit, causing negative feedback.

But either way, I prolly shoulda left it at, "it's likely a feedback winding, but I don't know how much feedback." 

[dscottguitars]

It looks like I have much more to learn about NFB.  I've used a pot in the past to adjust and liked that, but that was on a 6L6 50 watt amp which added a bit more NFB than the bassman to much less.  I may do that again, or use it to dial it into where I want, measure and then put in the corresponding resistors.  Thanks for all the input...

[PRR]

> NFB.  I've used a pot in the past to adjust

That works good when you take NFB to a low-level stage. The amount of NFB power is tiny, a pot can handle it, the waste is negligible.

There are arguments against taking NFB back over several stages. Stability problems increase. Every stage adds its own signature, which beats against all the distortion partials of all the other stages inside the loop.

I suspect Hammond liked the sound of the simple cathode NFB. On the one hand, it takes a lot of power to skew power-tube cathodes; OTOH as Mac showed it is possible to extract some power from the cathodes.

Taking NFB all the way back to a driver stage has always been MUCH more poular, for what that's worth. And much easier to tweak, which is important to voicing a guitar's amp/speaker combination.

[dscottguitars]

I think I know what you mean.  I put the NFB on the tail of the phase inverter as in this drawing of an amp I built, and the picture- I just finished it last weekend...