8 or 16 Ohm Tranny Tap?

[zozoe]

Hiya all on this very sad night~~ About speaker loads & trannies: Is there ANY difference in sound if I use a particular 16 ohm speaker using the 16 ohm tap, as opposed to me using the exact speaker, except in 8 ohms, going through the 8 ohm tap? I think I know what the consensus is, but is it true that using more tranny iron benefits the overall sound?
Thanks

[xtian]

To my ears, no. And I've heard some smart folks agree. That "using the whole winding" thing is nonsense.

[Stevem]

If you think that the whole transformer winding thing is nonsense, then take one of you guitars pickups and strip out half of the windings and post up how the tone and output has changed when you hook it all back up and play it!

Have you ever played a guitar with coil taps on the pickups, because it's the same deal!

If your driving a output transformer very hard then using the full secondary winding applys the current load and it heat over more windings making the transformers life span greater.

[martin manning]

If you think that the whole transformer winding thing is nonsense, then take one of you guitars pickups and strip out half of the windings and post up how the tone and output has changed when you hook it all back up and play it!

Have you ever played a guitar with coil taps on the pickups, because it's the same deal!

No, not the same thing. The turns are reduced, but the load is the same.

If your driving a output transformer very hard then using the full secondary winding applys the current load and it heat over more windings making the transformers life span greater.

The power dissipated on the transformer will be the same if the load matches the tap in use. Granted the heat generated will be distributed differently within the coil, but I don't know if that would significant or not. It would depend on the winding pattern.

How would you judge this anyway? At minimum I think you'd need two of the same make and model speakers with different impedances, and the speakers would still influence the result.

[roberto]

The example of the pickup is wrong, but please note what Sowter is saying on his website:

We normally provide a single 8 ohm secondary or 4 separate secondary windings which can be arranged for loads of 16, 8, 4 or 1 ohms as shown in the tables below. Alternatively we can provide a single 4, or 16 ohms or 8 tapped at 4 ohms. Also we can provide twin 4 ohm windings which can be connected in series for 16 ohms or in parallel for 4 ohms. Tapped secondary windings will have somewhat increased losses, reduced HF response and reduced damping when the load is connected to a tap. This is particularly true when a 4 ohm tap on a 16 ohm winding is used. Please be sure to specify your preferred option if you wish to have a non-standard secondary option or a screen tap %.

https://www.sowter.co.uk/push-pull-outp ... ormers.php

[pdf64]

The example of the pickup is wrong, but please note what Sowter is saying on his website...

But the context is that Sowter design/build primarily for hifi / studio, for which the response up at 20kHz and way beyond is crucial.
Whereas what goes on 'up there' is immaterial for a guitar rig.

[wpaulvogel]

I was told by Chris Merren that the lower impedance tap of a multi tap output transformer will be stiffer feeling and the high impedance tap a bit softer. He was referring to his Dagnall C1998 version and comparing the 4 ohm and 16 ohm taps. I didn’t really feel that it was different.

[zozoe]

Cool, Chris is an old friend of mine for 30 years, & did work for me when we both lived on the east cost. Swell guy & he surely knows his stuff... I think I'd go the 'softer' route
Thanks

[Travis_HY]

https://www.premierguitar.com/articles/ ... -they-dont

This article by the inimitable Steve Fryette explains some of the behavior that occurs with different speaker taps.

Personally, I like 16 ohms for pretty much everything. 8ohms is fine. 4ohms is really not ideal to me, but I understand it is an important aspect of the classic Fender designs.

Some guys are specific about it.

Scott Henderson prefers 8 ohm cabinets, Steve Stevens prefers 16 ohm cabinets, Dumble seemed to favor 4 ohm cabinets so it's personal taste, ultimately.

[mhuss]

Nice article, but it does not discuss much about output transformers.

A solid state amp acts like a classical voltage source, and will supply as much current as necessary (within its limits) to put the desired voltage across the load/speaker. This is why watts delivered go up when the impedance goes down. (100w/8ohms, 200w/4ohms, etc.)

A tube amp works based on a matched impedance principle, where the ideal condition is where the source and load impedances match. It will never deliver more than it's rated power, and will only deliver its rated power when the source and load are matched. The reason many OTs have multiple taps is for the practical reason of being able to deliver rated power into any convenient speaker box.

The other complication is reflected impedance. A transformer does not have an implicit impedance, it merely reflects the load it 'feels.' So, if the primary is 6k6, and the secondary is 16 into 16, everything works as designed. If you hook up an 8 ohm speaker to a 16 ohm tap, the impedance the output tubes 'see' goes down as well, the tubes get stressed, and the system no longer works as well. This is also true if you mismatch the 8 and 4 ohms taps.

As a previous poster noted, 'identical' speakers are not identical, due to piece-to-piece variations in magnets, cones, surrounds, spider, etc., so a true comparison of 8->8 vs. 16->16 is not really possible or 100% valid. The voice coils would be different at a minimum.

I've never tried a listening test, but the difference in resistance (which would affect damping/'squish') between the secondary winding taps is so small (<<1 ohm), I can't imagine it having a significant effect on the sound.

[pdf64]

...A tube amp works based on a matched impedance principle, where the ideal condition is where the source and load impedances match...

I think that's something of a myth
A power amp with triodes may turn out such that source and load impedances match, but power pentodes (both beam and screen grid types) have a much higher output impedance. Hence their native behavior is more akin to a current source. The application of NFB will tend to bring it down, may coincidentally turn out to be equal to the intended load impedance, but that's immaterial to the actual power output, more of an output damping thing.

[xtian]

Wow. Raise your hand if you're grateful you can build good amps by standing on the shoulders of giants.

[Travis_HY]

Nice article, but it does not discuss much about output transformers.

...I can't imagine it having a significant effect on the sound.

I shared that article because it is also a part of Stevie's "series" of articles about speakers which has some interesting information about speaker impedance and frequency response. It has some information about the frequency response changing when using different taps, but it isn't a definitive article about this specific subject. I apologize if that confused anyone.

Now whether this phenomenon is caused by the changes in negative feedback or the actual transformer taps themselves I do not know with my limited scientific knowledge. What I do know is that those taps do sonically affect the response of the amplifier in one way or another and other players hear the differences I hear as well. That's all!

I love your website! All the best!

[roberto]

I don't recall an amp having the same primary impedance and tube internal resistance.
At least not by purpose. Triodes have usually lower internal resistance than primary inductance.
Pentodes in instrument amps have usually higher internal resistance than primary inductance.

Then there are ways (ultralinear taps, cathode feedback, local feedback, global feedback, etc...) to reduce their internal resistance to even lower values than triodes.
Other than linearity, the purpose is damping (I've seen it up to 70).

[pdf64]

Sorry for the slight derail, but 'matching impedance for max power' thing is a widely held misapprehension. I raised a couple of queries on the topic of tube amp max power delivery / output impedance over on MEF a few years ago, and got some great info from Steve Conner, J M Fahey, Dave H, to name a few.
https://music-electronics-forum.com/for ... -impedance
https://music-electronics-forum.com/for ... -impedance

Steve Conner gave a great explanation

The output impedance of a triode does actually look like a resistance equal to the plate resistance, and the maximum power theorem applies.

Pentodes and beam tetrodes have a very high and non-linear output impedance that can't be used as the basis for an impedance matching calculation. To a first approximation, the output impedance is infinite, and that's why the voltage goes through the roof into an open circuit. (Stored inductive energy in the output transformer helps the voltage on its way to infinity.)

Non-linearity is the key to the whole thing. A non-linear device can have different characteristics depending on the amplitude of the signal you examine it with. The high impedance of pentodes is a small-signal property: a small change in plate voltage leads to a much smaller change in current, so we say the impedance is very high.

But on the other hand, you can define the large-signal impedance in terms of the plate supply voltage used, and the maximum current that the tube can safely handle.

This is one way of choosing load impedances for pentode amps. The other way is to search empirically for the impedance that gives lowest distortion, if low distortion is what you want.

I made some measurements of the output impedance of one of my amps,
I've now tried 2 methods on another amp, which uses p-p 6L6 into 4k2:8 ohm OT @ static VB+ of abpout 425V (GZ34 rectifier, 47uF reservoir cap), no global NFB.

Using the method in post#1 (run amp at low level 4Vrms into an open circuit, then load it until the output voltage drops to 2V), it comes to ~ 90ohms.

Then using 'compare output into 8 ohm with output into 7 ohms' method. The 7 ohms was achieved by using a 56 ohm resistor in parallel with the regular 8 ohm load.
I checked it at 3 signal levels:
4Vrms (both 6L6 conducting for full cycle) ~ 104 ohms
13Vrms (21 watts@8 ohms ~ half power) ~ 51 ohms
18Vrms (max clean output before hitting the rails) ~ 51 ohms (the 7 ohm load didn't result in saturation)
25Vrms (max square wave) ~ 25 ohms