Hey
@politoleo could we get a bit of a comparison between Fryette Power Station vs IR-Load?
So the main difference on paper is that the Fryette is a mono tube poweramp and the IR-Load is solid-state and stereo.
The goals with both products are the same:
- Attenuate a tube amp.
- Amplify a low power tube amp to louder levels.
- Work as a poweramp for modelers and preamps.
- Allow some form of direct recording from XLR outs.
Both allow adjusting the poweramp behavior with Volume/Presence/Depth controls.
But after that the differences start to add up.
Reactive load:
The Fryette allows adjusting how the connected amp behaves via its loadbox switches. This in my experience allows
the connected tube amp to behave more like if it was connected direct to the cab. I use these switches to match the load behavior to my different cabs, and this allows me to make the Bypass (Amp in -> Speaker out) and Operate (Amp in -> Loadbox -> Poweramp -> Speaker out) states to sound/feel similar.
The IR-Load in comparison seems to have the same loadbox behavior with any amp. Thus the behavior of the
connected tube amp to my understandin would remain the same as the reactive load is always the same.
Poweramp:
The Fryette being a tube poweramp, it will always behave like a tube poweramp (albeit a fairly neutral one) into whatever speakers are connected.
The IR-Load by comparison needs to emulate behaving like a tube amp by using the speaker impedance measurement and machine learning model.
If I have understood your posts on TGP correctly, the Reactance knob governs how much of the ML model behavior is applied.
- All the way down and it would be pretty much neutral. Is it more like a "neutral tube amp", or pure solid-state behavior?
- All the way up would behave like a Friedman poweramp. I assume this means similar to e.g a Friedman BE poweramp. Marshallish, middle of the road in how tight/loose it is etc.
Usage with different amps:
With the Fryette you try to match the behavior with the load switches and Presence/Depth.
How would the IR-Load behave with say a Tweed or Vox AC30? The Tweed or AC30 is pretty saggy when cranked, which is at odds with how the poweramp behaves. If the machine learning model represents a more Marshall style poweramp, and the loadbox maybe approximates say a 4x12 cab, how would this work out compared to running a Vox AC30 into 2x12 Celestion Blues for example?
Just noticed this! I answered a lot on the other forum so I'm going to copy and paste much:
LOAD
Others:
Single inductor = very off sounding, very popular, don't think anybody is using it anymore in new product, unless for some amps with integrated load.
Two stages = Current state of the art. Most of it have a very high Q on the low resonance peak, thin and high.
High pass curve also with fixed Q due to single series LR, so either it start too early or it starts too late and the rise doesn't reflect the behavior on the speaker.
Three stages = Depending on the design it can perfectly replicate the high pass variable Q. I believe Fractal and Reload II may use this as their curve looks very good.
With all we found that they replicate the measured impedance curve of the cabinet at idle (no BEMF)
IR-LOAD:
High resonance peak is very wide, similar to what you'd get combining 2 or 4 speakers in a cab, cause each speakaer resonance is slightly different than each other and never combine on a single peak. (Sound result = not artificially boomy)
The third stage allows for the right high pass shape, giving the right impedance from 700hz onward (Sound result = punchy and not strangely scooped)
The impedance curve is being replicated taking into account the back EMF current. So it's not going to be the same as you passively measure on a cabinet, rather the real one an amp sees while you play. (you can't measure it with a sine sweep, rather with high volume playtrack and ffts on current and voltage etc..) (Sound result = Punchy and clear)
A note about the back-emf: Being the load "reactive" but not really a speaker in a cabinet, only the inductors are providing a little tiny bit of it, but they can't react like a speaker as there is no pressure, so you can end up with infinite different impedance curve based on how hard the amp is pushing the cabinet and the kind of cabinet (and the kind of amp). We replicated the curve of a moderately high pushed cabinet with a Marshall amp.
POWERAMP
Even when the reactance knob is all the way to zero, the ML is still working, so it's not going to be an ultra flat poweramp as it doesn't feel right.
Rather it is similar to the Power Station but even more transparent. Dave have a modded power station he uses all the time that is very transparent and would almost disappear in A/B tests, but still sounds and feel great (and that was our reference point and improove from) unlike a fully transparent class D amp.
When reactance knob is all the way up you get a full Friedman/Marshall based poweramp, so very different from the Power Station in pure "amp" mode.
You can still get all the shades in between using the reactance knob, so with the reactance knob at 9 o clock you get something similar to a production power station etc..
Depth and Presence works all the time.
USAGE WITH DIFFERENT AMPS
We tested with several different amps (including VOX style, it was also at the show), and almost everytime the reactance was sounding and feeling best at zero or very close to, while presence and depth could give you a granular control (still over reactance, as it's cabinet based, not a simple EQ) to do fine adjustments, most of the time it was 8/9 o'clock for both.
The ML models with reactance at zero will be much more linear, think about a phase inverter with a lot of feedback, so you won't get much tonal character, rather an hi-fi-ish tube amp.
![[Nathan]](/data/avatars/m/0/613.jpg?1726885163){kind=avatar}
