True but I was speaking purely of the perspective of the amp itself with no other variables. IMO if the amp sounds different into a load it is designed incorrectly.
Of course as has been said unless you want an "effects box" with a variable output impedance, in which case it will sound infinitely variable into different loads so trying to compare the amp to another is pointless to begin with.
The reason I made the thread was because I wanted to know if tubes themselves could add some magic to the sound. I think it's pretty obvious that they will sound like "something" when used in an output stage due to their imperfections under load but that is a given. But I wanted to know if they could be used to "spice" up the the sound if inserted into X circuit.
Last edited:
It is not particularly hard to come up with some first hand experience on this subject. All you need is a unity gain tube based stage you can insert into an existing amplification chain.
Basically 3 choices: a cathode follower, an amplification stage followed by a step-down transformer and an amplification stage preceded by an attenuator.
Basically 3 choices: a cathode follower, an amplification stage followed by a step-down transformer and an amplification stage preceded by an attenuator.
Several excellent solutions exist for tube sound conversion. The tube in fb of op-amp is my favorite. (sometimes called the super triode connection)
On straight wire vs tone controls, I'm installing a 'loudness' switch on latest board but on *rear* panel out of view. No one will know.
John Wayne liked the Vespa but did not let anyone see him riding it.
Don't ask don't tell method...
https://www.diyaudio.com/forums/tub...ode-connected-stc-chip-amp-8.html#post6031721
-
On straight wire vs tone controls, I'm installing a 'loudness' switch on latest board but on *rear* panel out of view. No one will know.
John Wayne liked the Vespa but did not let anyone see him riding it.
Don't ask don't tell method...
https://www.diyaudio.com/forums/tub...ode-connected-stc-chip-amp-8.html#post6031721
-
Attachments
this board after all is a diy'er board, where members are free to experiment and even make fools of themselves.....😀
i tend to look at the message and if there is nothing there move on and try again...
i used a bridged ACA amp driving a 15 inch 4 ohm Klipschorn in a triamp system, the other two amps being the 2a3 set and the 6b4g set amps...
bass was so realistic and loud as if we were hearing the real thing, kick drums and bass guitar riffs...loud but not irritating nor fatiguing to the ear...
but still the delineation between what a current amp and a voltage amp remains vague to me...
when does an amp become a current amp and when does it become a voltage amp, that is still a mystery to me..
I'm not Dave, but you might be interested in this:
The Pentode
It's about why 1933 radio designers liked the high output impedance of a pentode output stage, among other things.
Last edited by a moderator:
@TonyTecson - "when does an amp become a current amp and when does it become a voltage amp"
Earlier in the thread someone said (regarding tubes) "It (tube sound character) depends how they're used". Same with amplification elements in circuits rendering a voltage or current output.
Generally, a current source has high output impedance (up to infinity) while a voltage source has low output impedance (down to zero). Any single tube or transistor amplification stage generally has a high impedance output (Plate or Collector/Drain) and a low impedance output (Cathode or Emitter/Source); and a circuit can be arranged to emphasize one or the other.
If you look at the schematic of a transistor amp with low output impedance, you'll typically see transistor emitters or sources connected together to provide the output in a symmetrical push pull arrangement. These can be flipped so the collectors or drains are connected to provide the output. When I see this, I assume it's done to get the amp to have a higher output impedance than in the more "normal" common-emitter arrangement.
Tubes on the other hand are almost always used with the plate (high impedance) side driving a transformer to convert down to a lower impedance for driving the speaker. These lower impedances are typically your "4-8-16" Ohm taps on the amplifier. Not quite all the way to the Zero Ohm characteristic of an optimal voltage amp, but close. So a typical tube power amp output can be thought of as being somewhere between a voltage source and a current source output, as the output impedance is neither zero nor infinity; it's a number like "4". I'm sure that clarifies everything...
"Well why not just stick a non-inductive resistor in series with a voltage amp's 'hammer hard' common source output to make it also have an output impedance like "4"?" You can. It's just not as elegant as designing the circuit to render that property in the first place. But it will make the voltage amp sound a little different. One thing that would do is shift the driver Q up, to whatever effect that has on the driver/cabinet combo. This could be perceived as a little less "control" on the bass, or it could even make it sound better. Depends.
Hope this helps!
Earlier in the thread someone said (regarding tubes) "It (tube sound character) depends how they're used". Same with amplification elements in circuits rendering a voltage or current output.
Generally, a current source has high output impedance (up to infinity) while a voltage source has low output impedance (down to zero). Any single tube or transistor amplification stage generally has a high impedance output (Plate or Collector/Drain) and a low impedance output (Cathode or Emitter/Source); and a circuit can be arranged to emphasize one or the other.
If you look at the schematic of a transistor amp with low output impedance, you'll typically see transistor emitters or sources connected together to provide the output in a symmetrical push pull arrangement. These can be flipped so the collectors or drains are connected to provide the output. When I see this, I assume it's done to get the amp to have a higher output impedance than in the more "normal" common-emitter arrangement.
Tubes on the other hand are almost always used with the plate (high impedance) side driving a transformer to convert down to a lower impedance for driving the speaker. These lower impedances are typically your "4-8-16" Ohm taps on the amplifier. Not quite all the way to the Zero Ohm characteristic of an optimal voltage amp, but close. So a typical tube power amp output can be thought of as being somewhere between a voltage source and a current source output, as the output impedance is neither zero nor infinity; it's a number like "4". I'm sure that clarifies everything...
"Well why not just stick a non-inductive resistor in series with a voltage amp's 'hammer hard' common source output to make it also have an output impedance like "4"?" You can. It's just not as elegant as designing the circuit to render that property in the first place. But it will make the voltage amp sound a little different. One thing that would do is shift the driver Q up, to whatever effect that has on the driver/cabinet combo. This could be perceived as a little less "control" on the bass, or it could even make it sound better. Depends.
Hope this helps!
There are also articles by Philips describing the advantages of using 400 ohm loudspeakers (Philips produced hifi units up to 1200 ohm in the past). I had some 5K/400R SE outputs made for a pair of 400 ohm wide band units from salvaged 50s radios. In an OB of 50x60cm reproduction was jawdropping precise, running just from a pair of EL84s. Lowest freq would be 100Hz I presume but that problem can be dealt with. So to say there's more to be experimented with than 13-in-a-dozen equipment.
I have plenty in this context, as I've stated in the first page.
None of my experiments with tubes have resulted in anything other than a varying level of objective sound quality. No special sound.
Indeed, this is what I had planned. I have a few super triode variations of my own I wanted to build but there is literally no point if tubes don't "sound" like anything better than just making a good SS amp.
I might add - long as I'm blathering - that a low Z amp output is preferred by engineering types. It's like putting a solid stake in the ground as reference. The speaker will respond to a series of similar lo Z voltage amps in the same way. Makes for a good consistent product - assuming all amps used are the common, low Z output voltage type.
Of course, the same could be done assuming all amps used are going to be tube types with an output Z something like "2". However variations about the value "2" would be greater across tube amp types, than variations about "0.01" across transistorized voltage amps - and have more effect; 1.5 > 2.5 presenting greater effect than 0.001 > 0.05 Ohm. With 4 - 8 ohms speakers.
The quest for very low Z output voltage amplification has been going on long before I was born, and I consider myself to be an old dude these days. I stumbled upon this while researching R-J speaker cabinets -
Of course, the same could be done assuming all amps used are going to be tube types with an output Z something like "2". However variations about the value "2" would be greater across tube amp types, than variations about "0.01" across transistorized voltage amps - and have more effect; 1.5 > 2.5 presenting greater effect than 0.001 > 0.05 Ohm. With 4 - 8 ohms speakers.
The quest for very low Z output voltage amplification has been going on long before I was born, and I consider myself to be an old dude these days. I stumbled upon this while researching R-J speaker cabinets -
Attachments
Last edited:
that 1953 amp article is interesting (wonder how it might sound on a horn system ?) - it continued on page 67 rather than 66:
https://i.imgur.com/J5E6PMb.jpg
I like the old R-J approach and had 4-18" speaker cabinets vented that way - the originals are cool too other than cabinet too thin and no bracing. JBL's diffraction slot might make a fun modern version vs R-J's classic "lemon" aperture.
https://i.imgur.com/J5E6PMb.jpg
I like the old R-J approach and had 4-18" speaker cabinets vented that way - the originals are cool too other than cabinet too thin and no bracing. JBL's diffraction slot might make a fun modern version vs R-J's classic "lemon" aperture.
But many of those unwanted variables need to be in place before you can properly test amplifiers. The lack would make headphones not representitive of the load they have to drive so you may seem to have greater “clarity" using headphones but they ar enot properly stressing the amplifier.
dave
I wish I could remember where I read it (here perhaps?), someone wrote an article about relatively high second harmonic distortion in SE amps, and if it was matched with a similar distortion full range speaker, anti phase, the distortions would cancel giving a very low distortion acoustic output. Multiway speakers confuse the issue, but the principle remains. Does anyone else remember this?
Without those variables, the test is not valid. An amplifier should always be considered as a system with the speakers it drives and the cable that connects them.
Now if you have a typical set of speakers with an “ugly” impedamce curve you need a low output impedance amp. With a world of only gear within that subset the issues with trying to test amps (or speakers) independently then many of the variables go away, but as soon as you start delving into the fringe that goes out the window.
dave
And the imperfections of SS devices. Unlike SIT-like devices transistors & MOSFETs do not have near as linear a transfer function as a good triode. So immediately SS is at a disadvantage (in most cases).
Now a lot of tricks have been tried to make typical SS devices act more linearly — the one that comes to mind id the Scade feedback amp Nelson outlined at BAF a couple years ago.
Execution is critical. The device itself is not a determiner. But it is typically simplier to understand & execute a tube amp that is good.
dave
When bridged, each channel sees half the load, so here each channel sees 2Ω. The ACA starts struggling as you approach 4Ω so this situation is likely not getting the best out of the amp with that speaker. I suspect parallel operation (each amp sees 8Ω) would give you better performance.
Well officially the dividing line is Rout> Rspeaker. A current amp will convolve part of the impedance curve with the FR. ie if you have a bump in the impedance, you will get a bump in the actual FR. If this compensates for a dip (when measured with a low output impedance amp) things are good, but hoping for that is a bit of a crapshoot because it is not a stark transition but a slow transition from one to the other.
A current amp needs a speaker with flattish impedance such that the interaction with the speaker impedance is moot or the “effect” benefits the end performance of the loudspeaker.
The interesting stuff where one has to be even more consious of the interaction are those situations where Rout is near Rspeaker, the most common amp falling into this set id the single ended triode amp, and (to my mind) the most significant SS representitive, the Firstwatt F2 (the F1 has a high enuff output impedance to be considered a current amp).
dave
I was going to mention the same. The URL is long lost but a fellow in Europe did a bunch of tests with a SET, a typical Class AB SS amp (NAD) and a set of big horns with FE206e.
The independent mesaures of the amp's THD gave the NAD a huge edge, but when both amps were connected to the speaker, the SET/Speaker combo gave lower measured distrortion than the NAD/speaker combo. (and sounded better — the FFxx6 series is intended for use with typical SETs).
Now you have to put some stock in (single number) THD measures but it does show the importance of “system”.
dave
It is only useful if it has exactly the load of the speaker you are going to use, why not just use the speaker.
Headphones just don’t cut it — unles syou are testing headphone amps, and with headphones there could well be even greater issues with system matching given the huge variability of headphone impedances out there.
dave
- Status
- Not open for further replies.