"Fast amp" vs "Slow amp", What's the difference?

[sjturbo]

I asked this on another forum but it's not strictly amp oriented, so I thought I would ask here.  I've heard some amps are considered "Fast" (AC30) and some "Slow" (5E3). I assume this is based on the circuit design? I don't know if I can wrap my mind around the answers I may get but would appreciate the feedback!

[TIMBO]

This is only my 2c. I would consider that a amp that has a high voltage B+ and large filter caps and SS rectifier to be a fast amp and an amp that has low B+ voltage with small filters and a tube rectifier that has lots of sag to be slow. 

[John]

From tidbits I've read here and there I think Timbo is right... Eric Johnson, for instance, probably doesn't want a lot of sag, where a guy that plays slow, like... me... likes a little sag as it lends character to my slow playing. 

[HotBluePlates]

I've heard some amps are considered "Fast" (AC30) and some "Slow" (5E3). I assume this is based on the circuit design?

I've only heard 1 person* use this terminology for an amp, where it might have been technically accurate. For the most part, these terms cropped up in magazine articles, and are based on the writer's subjective opinion on the amp's response.

My guess is that the terms are applied based on the amps' treble response, or perceived lack of sag.

Amps with a solid power supply or a large output power give the impression of a fast response. I used to own a Fender Tonemaster amp; the Vintage 30 speakers (which do not sound like vintage Celestion 30w speakers) and the four 6L6's made it sound stiff and bright when playing at home or at low volume. But when I played onstage, people commented how great the amp sounded, probably because it cut through well without being terribly "zingy".

Anyway, some writers will give you the impression a "fast amp" is what you want. However, it really depends on what sound you're trying to achieve and what works for the style of music you're playing. Since we only hear up to ~20kHz (and there's no useful guitar signal or harmonic that high), and our speakers don't respond that high, I'd argue a "fast amp" by a technical definition is not useful to us.
* That guy was Dave Funk, when referring to some amps he built in maybe the early-90's under the name "Triggs". The amps used "linear crystal wire" (whatever that is), and he mentioned the response was very fast.

Dave has some serious electronic knowledge. To folks with a technical background, "fast response" is often associated with how high a frequency a circuit can pass. This is tested with square waves, and a fast circuit (good HF response) will pass a square wave without rounding the edges of the wave. However, you usually look at such things when moving into radio frequency. By this standard, a 100mHz scope is a "faster circuit" than a 10MHz scope. The circuit with higher high frequency response can follow a faster-changing (higher frequency) signal better.

Put yet another way, a "faster change" occurs in a shorter time, and since frequency is the inverse of time, a shorter time implies a higher frequency.

[RicharD]

I thought this was a solid state term used to express the slew rate of the output chip.

[HotBluePlates]

Yeah, slew rate is another "proper application" of the term "fast".

Still amounts to the same thing (though the reasons are different): how high (i.e. how "fast") can you go?

But slew rate is also about how big an output you can get at that high frequency, and goes a bit beyond this discussion (cause it mostly applies to solid-state).

[FYL]

Yeah, slew rate is another "proper application" of the term "fast".

Still amounts to the same thing (though the reasons are different): how high (i.e. how "fast") can you go?

But slew rate is also about how big an output you can get at that high frequency, and goes a bit beyond this discussion (cause it mostly applies to solid-state).

Yup. Slew rate is usually expressed as V/µs, volts per microsecond. It represents the max rate of change of a signal. Rise time (in µs) is related, but only deals with bandwidth.

All MI tube amps I'm familiar with - there could be a couple of OTL amps in labs somewhere - use output transformers and show a limited bandwidth, so their rise times are quite high. As the signal is very limited in bandwidth - say 6 KHz or so for typical pups - this isn't a problem.

Perceived "speed"  IMO relates to high transconductance or paralleled tubes, a forward balance and/or tight supplies.

[jjasilli]

IMHO:  sag, as mentioned; and choke.  A resistor is fast; a choke woks by induction which works by build-up, then dissipation of a magnetic filed.  These things occur over time, and hence lag behind signal, making the amp seem "slow".

[moonbird]

I am NO expert --

But if I remember reading KO correctly, a class A amp cannot SAG reguardless of the other components involved. A class AB amp can have sag, but a choke acts to combat sag by resisting change to current flow once the power-up activity has completed. A choke makes the PS stiffer.

Guess I am wondering at this point if 'sag' is the same thing as 'slow'. Seems to be confusion.

[Tone Junkie]

Maybe Im out in left field and this might have nothing to do with it . but to me a fast amp would be when Im playing a bit of heavy music i consider a fast amp one that can take fast palm muting type riffs with a good bit of over drive and be clear and precise were a slow amp is all jumbled or distorted without picking clarity with the same amount of over drive or distortion.

So my defanition is slow amp muddy without clarity, fast amp is articulate with defanition while both having the same amount of gain. both being in high gain territory.
Thanks Bill                 

[jjasilli]

I'm thinking that a screen choke will not make the plate supply stiffer; and the choke still lags, especially for transient signal spikes, because of the time needed for build-up and then persistence of the magnetic field. 

In Class A the plate is already drawing full current, so there's less opportunity for sag. Whether sag can happen in Class A seems to be controversial.  I need to pull my Vibrochamp chassis (I added a screen choke), and do some voltage readings under various signal conditions.

[moonbird]

I'm thinking that a screen choke will not make the plate supply stiffer;

Good point ... but it would still have some impact on output right?

I need to pull my Vibrochamp chassis (I added a screen choke), and do some voltage readings under various signal conditions.

Wow -- that would be GREAT data to see!!  FWIW I DO NOT think this is a settled issue either.

[moonbird]

Thinking a bit more --

It seems that there is marketing hype out there that says that *ole skool* electrolytic PS caps are "slow" and that very expensive and large polyprop PS caps are faster. Been meaning to test that one myself -- but as I am sinking down into the easy chair in front of the boob tube I always get amnesia about these things I *need* to do seems like.    .

Would be VERY interested in other's experience with these polyprop PS caps..

[HotBluePlates]

But if I remember reading KO correctly, a class A amp cannot SAG reguardless of the other components involved. A class AB amp can have sag...

A class A amp is not supposed to sag due to output tube plate current draw.

Theoretically, the average current drawn by a class A output stage is equal to the idle current draw; no net change in current yields no sag through the power supply. In reality, depending on a number of factors, there may be rectification effects exhibited by the output tube(s), which results in a net higher current drawn at full power than at idle. If you peruse tube data sheets, the "Class A typical operation" will bear this fact out; idle current is generally at least a little lower than full power current.

You can also force sag to occur, regardless of variation of plate current, if there is appreciable screen current and you use an oversize screen resistor. You won't see something like this except in a homebrew amp, where the builder wanted to add sag where it might not otherwise occur.

I'm thinking that a screen choke will not make the plate supply stiffer; and the choke still lags, especially for transient signal spikes, because of the time needed for build-up and then persistence of the magnetic field.

I'm with moonbird on this. A choke makes the power supply closer to ideal.

Think about it for a second: the alternative to a choke is a series resistor; you need the series resistor to be a fairly big impedance compared to the filter cap to get a reduction in ripple. A choke has a fairly small d.c. resistance compared to the reactance at the ripple frequency. Obviously, the choke's d.c. resistance will cause much less variation in output voltage with a changing current draw.

I think you are basically right about how a choke works. But don't forget that it is the inverse of a cap in almost every way. Where a chokes tends to release energy to keep a current steady, a cap releases energy to keep a voltage steady. The "lag" that you mention is a non-issue in this situation; with a choke current lags voltage, but with a cap voltage lags current. A "fast amp" can be built using either (or both).

Maybe Im out in left field and this might have nothing to do with it . but to me a fast amp would be when Im playing a bit of heavy music i consider a fast amp one that can take fast palm muting type riffs with a good bit of over drive and be clear and precise were a slow amp is all jumbled or distorted without picking clarity with the same amount of over drive or distortion.

And that's how most people think of it.

But think about this: isn't what you describe just a solid power amp, with "right EQ" to get rid of mud?

It seems that there is marketing hype out there that says that *ole skool* electrolytic PS caps are "slow" and that very expensive and large polyprop PS caps are faster.

Would be VERY interested in other's experience with these polyprop PS caps..

This isn't a good reason to use polypropylene caps.

Any polypropylene cap will be closer to an "ideal cap" than any electrolytic. That just means that there are fewer and smaller parasitic artifacts (lower dissipation factor, less leakage, less variability with voltage/temperature, less variation of capacitance with frequency, etc). Polypro's will also have tighter tolerances and a very much longer lifespan than e'lytics.

I've used them in a tweed Deluxe copy. The amp sounds like a tweed Deluxe. The only reason I used them is that I'd likely never have to replace a filter cap.

[moonbird]

Hot Blue Plates --

Thanks for the comprehensive post and the reminder about Class A and sag. Indeed there is generally a 2-3 ma difference in plate current on most tubes in Class A. Never thought of the significance of that before. Seems to differ depending on configuration though. Consider the following data from NJ7P web site @ 250V plate voltage - for 6v6gt tubes in class A:

         Plate Current (Zero Signal) ...................    45 mA
         Plate Current (Maximum Signal) ................ 47 mA
         Grid No. 2 Current (Zero Signal) ..............    4.5 mA
         Grid No. 2 Current (Maximum Signal) ........... 7 mA

Whereas for 6v6gt @250V Class AB1:

         Plate Current (Zero Signal) ...................    70 mA
         Plate Current (Maximum Signal) ................ 79 mA
         Grid No. 2 Current (Zero Signal) ..............    5 mA
         Grid No. 2 Current (Maximum Signal) .......... 13 mA

@285V Class AB1 it is even better (or worse):

         Plate Current (Zero Signal) ...................    70 mA
         Plate Current (Maximum Signal) ................ 92 mA
         Grid No. 2 Current (Zero Signal) ..............    4 mA
         Grid No. 2 Current (Maximum Signal) ......... 13.5 mA

You can also force sag to occur, regardless of variation of plate current, if there is appreciable screen current and you use an oversize screen resistor. You won't see something like this except in a homebrew amp, where the builder wanted to add sag where it might not otherwise occur.

This point is very interesting to me. Can you site an example of a design like this that I can look at? What tube types or tube characteristics should one look for to best accomplish this? I am most interested in blues playing and low-output, studio type amps so even small signal pentodes would do for me most likely.

Lastly -- thanks for the report on the polyprop caps -- been wanting to use them -- but afraid of losing the "mojo". Should have known better. Thanks again.

[jjasilli]

So for Class A, there seems to be a total of 5.5mA more current draw from idle to full signal. Overdrive may produce more current draw. But I'm not sure what resistance figure to use to compute a DC voltage drop in the B+ power supply.   

For screen resistor Fender used none, or 470R.  Marshall uses 1K or 1.5K 5W.  KOC recommends 2.2K 3W for PP EL84's.  I tried that once but didn't like the tone -- but other factors were then involved, such as very high plate voltage over 400VDC, so I wouldn't judge KOC on that one anecdotal experience.  Anyway for big bottle tubes 5W resistors are recommended. 

[moonbird]

On the NJ7P site these values all presume that plate and screen voltage is the same 250V or 285V respectively.

[jjasilli]

Yes , but his must be at idle; i.e. no signal, condition.  As signal increases to max signal, this changes.  Also guitar amps exceed tube spec max signal to purposely get distortion, or overdrive.  This was not contemplated in the tube specs.

Hotblue's response has me surmising that in the hi-fi world regulated power supplies are desirable to keep the power amp from deviating from the average B+ condition -- i.e. from wavering over time, being "slow", and affecting the pure injected signal.

[HotBluePlates]

Consider the following data from NJ7P web site @ 250V plate voltage - for 6v6gt tubes in class A:

         Plate Current (Zero Signal) ...................    45 mA
         Plate Current (Maximum Signal) ................ 47 mA

Well, let's look at a 6L6GC data sheet. Class A, 270v plate and screen, 2 tubes, we go from 134mA at idle to 155mA at maximum signal. That's 2 tubes, so each tube is moving from 67mA up to ~77mA. This is not the peak current through each tube, but the average. Where we think the current increases during part of the cycle, then decreases, and it all averages to the same thing as the idle value, well, that's a little bit removed from reality. If the tubes were perfectly linear at all signal levels, then that would be true.

Anyway, you can read about this stuff in RDH4 (and other old tube books), where it's talked about as "rectification effect" or a "shift in loadline".

Can you site an example of a design like this that I can look at? What tube types or tube characteristics should one look for to best accomplish this?

Tubenit has used this often in his amps. It was a suggestion I made to him some time ago, and it seemed to work well for the type of music he plays.

Honestly, I don't think the effect was calculated or designed, but trial and error to see what sounded best. Sag by way of "too big" screen circuit resistance is something that's we've talked about a lot on this forum. It's something to use judiciously, because it does rob an amp of power output it would otherwise be capable of producing. Sag from any source generally takes place at the point the amp stops putting out increased power and moving to increased distortion instead.

Of course, O'Connor has a Sag Control Kit, which adds a variable impedance in series with the B+ to force sag at any volume level. It's kinda like power scaling, but not, in that the variable impedance is a MOSFET controlled by a small control circuit and a pot.